DEV Community: Quadratic

Database analytics: Query your source of truth

Quadratic — Wed, 15 Oct 2025 20:34:08 +0000

Product managers have to solve a fundamental database analytics problem in today's data-driven business environments. You need fast, reliable insights to validate experiments, measure feature impact, and guide product decisions. However, the traditional path to those insights often involves waiting days for analyst availability, navigating rigid business intelligence dashboards, spending long amounts of time manually analyzing the data, or getting blocked entirely from database access by well-intentioned but overprotective data or IT teams.

These data analysis bottlenecks affect how product managers approach experimentation and measurement. For example, when you cannot get answers quickly, you may skip experiments altogether or make decisions with incomplete information rather than manually do complex analyses. Both outcomes undermine the data-driven product development that modern companies depend on.

In contrast, there is a new approach to database analytics focused on data accessibility and democratization. The new approach, led by tools like Quadratic, includes an AI embedded in the tool with a direct connection to your database. This closeness to data puts product managers in direct conversation with their data sources through natural language interactions with the AI. This shift from request-based to self-service analytics transforms the speed of insights and makes it easy to increase the depth and frequency of product experimentation.

Where the traditional process breaks down

Even when data or IT teams quickly respond to requests, you receive a static report that may not answer your questions completely nor provide answers to your follow-up questions. Thus, the process breaks down when product managers need to iterate quickly on hypotheses, explore data interactively, or adjust their analysis based on initial findings. The lag between question and answer kills the momentum that drives effective product experimentation.

Database access restrictions compound this problem. Many organizations limit direct database access to protect data integrity and prevent accidental performance impacts. While these concerns are valid, they often create an overcorrection that blocks product managers from the very data they need to make informed decisions about their products.

Consequently, product managers, who are closest to user problems and product opportunities, have the least direct access to user data. This separation between decision-makers and data creates a systematic delay in product learning that accumulates over time, slowing down the entire product development cycle. This can lead to an actual breakdown of the process when product managers decide it is too difficult to get the desired information.

Choosing the right database and analytics approach

The foundation of effective database analytics lies in selecting the combination of databases and analytical tools that match your team's technical capabilities and business requirements. The question of which database is best for analytics does not have a universal answer. Instead, understanding the key considerations helps product managers make informed decisions about their analytical infrastructure.

Database and analytics integration requires careful consideration of performance, scalability, and ease of use. While some teams benefit from dedicated analytical databases optimized for complex queries, others find that their existing production databases provide sufficient performance for product management use cases. The best database for data analytics often depends more on your team's existing infrastructure and expertise than on raw technical specifications.

In-database analytics represents an increasingly popular approach that aligns with modern ELT (Extract, Load, Transform) architecture rather than traditional ETL (Extract, Transform, Load) pipelines. Instead of extracting data for external transformation and analysis, in-database analytics follows the ELT pattern by loading raw data into your database first, then performing transformations and analytical calculations directly within the database environment. This approach significantly reduces data movement overhead, eliminates the complexity of external transformation steps, and improves query performance for complex analytical workloads.

Modern database analytics tools provide abstraction layers that make databases in data analytics more accessible to product managers without extensive SQL expertise. These tools often include query builders, visualization capabilities, and collaborative features that bridge the gap between technical database access and user-friendly analytical interfaces.

Understanding modern database analytics vs traditional BI

Traditional business intelligence (BI) tools excel at providing consistent, governed views of key metrics across an organization. They offer standardized definitions, reliable refresh schedules, and user-friendly interfaces that make complex data accessible to non-technical stakeholders. However, this strength becomes a weakness when you need to explore data in ways that were not anticipated by the dashboard designers.

Database analytics represents a fundamentally different approach to working with data. Instead of consuming pre-aggregated reports or navigating through predetermined dashboard views, database analytics involves querying your source of truth data directly to answer specific questions as they arise.

When you run an experiment that involves user segments not captured in existing reports, or when you need to blend data from multiple sources to understand a conversion funnel, traditional BI tools often fall short. You end up exporting data to spreadsheets, manually joining datasets, and losing the governance and accuracy that made BI tools valuable in the first place.

In contrast, the best database query tools provide direct access to source data while including the analytical capabilities provided by the AI. Rather than being limited to predefined reports, you can write SQL queries that precisely target your analysis needs, join data across multiple tables, and iterate on your approach until you find the insights that drive decisions.

This approach particularly benefits product managers because product decisions often require custom analyses that do not fit standard reporting templates. When you're evaluating whether a new onboarding flow improves activation rates, you need to define cohorts, track user journeys, and measure outcomes in ways that are specific to your product and user base. This type of analysis of the database requires flexible database analysis tools that easily adapt to your specific analytical needs.

The SQL advantage for product experimentation

SQL data analysis tools provide product managers with a powerful language for asking precise questions about user behavior and product performance. Unlike point-and-click interfaces that limit you to predetermined analysis paths, SQL allows you to express exactly what you want to measure and how you want to segment your data.

For product managers running experiments, this precision is crucial. When you launch an A/B test, you need to ensure that your analysis accounts for factors like user segments, time periods, and statistical significance. Writing SQL queries allows you to explicitly define these parameters rather than hoping that a dashboard or analyst has anticipated your needs.

For someone comfortable with spreadsheet formulas, the typical learning curve to do meaningful analyses in SQL may be a few weeks. In contrast, in Quadratic, the learning occurs in minutes. You write and refine queries by describing what you want in natural language to the AI. It then writes the SQL to query your data.

When you have the AI write your SQL queries, you maintain complete visibility into how your results are calculated. This transparency is essential when presenting findings to stakeholders or when you need to modify your analysis based on new questions. Rather than treating your analytics as a black box, you can explain exactly how you arrived at your conclusions or, if needed, you can even show the code that was used.

Database analysis approaches and techniques

Effective database data analysis encompasses multiple analytical approaches that serve different product management needs. Retrospective database analysis allows product managers to understand historical patterns and learn from past product decisions. This approach involves examining user behavior data, experiment results, and feature adoption patterns to identify trends that inform future product strategy.

Semantic database analysis focuses on understanding the meaning and relationships within your data structures. This technique becomes particularly valuable when working with complex analytics database schema that include multiple interconnected tables representing users, events, experiments, and business outcomes. Understanding these semantic relationships enables more sophisticated SQL database analysis to answer complex product questions, and semantic layers can play a huge role in improving AI understanding of database data.

The choice of database analysis software significantly impacts your analytical capabilities. While some teams prefer free database analysis tools that provide basic query functionality, others invest in comprehensive database visualization tools that include collaboration, Python support, and database analysis & reporting capabilities. The best SQL software for data analysis often combines query flexibility with user-friendly interfaces that accelerate insight generation.

Creating standardized database analysis report templates helps ensure consistency across different product investigations while reducing the time required to generate insights. These templates typically include common database analysis techniques like cohort segmentation, conversion measurement, and trend analysis that can be adapted for different product questions.

Building repeatable analysis workflows in database analytics

One of the most powerful aspects of database analytics is the ability to create analysis templates that can be reused and adapted for multiple experiments and investigations. Rather than starting from scratch each time you need to measure something, you can build a foundation of core queries that pull the metrics most relevant to your product decisions.

This approach transforms ad hoc reporting and analysis from a time-consuming custom project into a rapid exploration process. You can define queries that connect to your key data sources, which may be Mixpanel events, UTM tracking data, Stripe revenue information, user behavior logs, or something else. Then you can quickly generate new insights by simply modifying parameters or adding filtering conditions.

This foundation-building process typically involves identifying the core data entities that drive your product decisions. For most product managers, this includes user data, event tracking, experiment assignments, and revenue metrics. By creating reliable queries that join these data sources, you establish an analytical base that can support multiple types of investigation.

Once this foundation exists, generating new analyses becomes significantly faster. Instead of recreating the data pipeline for each experiment or feature evaluation, you can focus your time on interpreting results and determining next steps. This efficiency gain is particularly valuable during intensive experimentation periods when you might be running multiple tests simultaneously.

Practical database analytics workflows for product managers

Quadratic’s most effective database analytics workflows for product managers combine the precision of SQL queries with the analytical power of programming languages like Python. This combination allows you to extract exactly the data you need from your database and then perform sophisticated analysis and visualization without switching between multiple tools.

A typical workflow might begin with a SQL query that pulls user data, experiment assignments, and outcome metrics from your database. This query serves as your data foundation, ensuring that you are working with accurate, up-to-date information that reflects the current state of your product and users.

Then you can use AI-powered analysis to quickly generate insights and visualizations. Rather than manually calculating conversion rates or creating charts, you can describe what you want to understand in natural language, and the AI writes the code that performs the analysis and generates the charts. This approach maintains the transparency and reproducibility of written code while dramatically reducing the time required to generate insights and visualizations.

The key advantage of this workflow is its adaptability. When stakeholders ask follow-up questions or when you discover unexpected patterns in your data, you can quickly modify your analysis without starting over. This flexibility is crucial for product managers who need to iterate on their understanding as they learn more about user behavior and feature performance.

Overcoming database access and governance challenges

Many product managers encounter organizational resistance when requesting direct database access. Data teams often worry about query performance, data security, and maintaining consistency across different analyses. These concerns are legitimate, but they can be addressed through thoughtful approaches to database analytics that balance access with governance.

Modern analytical databases provide multiple mechanisms for granting controlled access to production data. Rather than giving product managers unlimited access to primary databases, organizations can create read-only replicas, implement query timeout limits, and establish clear guidelines for responsible data use.

Cloud-based analytics database solutions often include built-in governance features that make controlled access easier to implement. The best database for analytics data access typically includes role-based permissions and query monitoring capabilities.

The governance challenge can also be addressed through collaborative approaches where data teams help product managers establish their analytical foundations while maintaining oversight of query patterns and resource usage. This collaboration often results in better outcomes than either complete restriction or unlimited access.

Documentation plays a crucial role in successful database analytics governance. When product managers document their queries, analysis methods, and key findings, they create institutional knowledge that benefits the entire organization. This documentation also helps data teams understand how product decisions are made and where additional support or optimization might be valuable.

Measuring the impact of direct database access

Organizations that successfully implement database analytics for product managers typically see improvements in both the speed and quality of product decisions. The most obvious benefit is reduced time-to-insight, with many teams reporting that experiments can be analyzed within hours rather than days or weeks.

However, the more significant impact often comes from increased experimentation frequency. When product managers know they can quickly analyze results, they're more likely to run smaller, more targeted experiments that provide focused learning. This shift toward continuous experimentation accelerates product learning and leads to more informed product decisions.

The quality of analysis also tends to improve when product managers have direct access to data. Rather than relying on generic reports that may not capture the nuances of their specific experiments, they can create custom analyses that precisely measure what matters for their product decisions.

Teams should track metrics like time from experiment completion to decision, frequency of data-driven product changes, and stakeholder confidence in analytical results to measure the impact of implementing database analytics workflows.

Advanced techniques for product analytics

As product managers become more comfortable with database analytics, they can leverage advanced techniques that provide deeper insights into user behavior and product performance. Cohort analysis becomes much more powerful when you can dynamically define cohorts based on user actions, experiment participation, or product usage patterns.

Funnel analysis gains precision when you can have the AI write custom SQL that accounts for the specific user journeys relevant to your product. Rather than relying on predefined funnels that may not match your actual user experience, you can create dynamic funnels that adapt to different user segments or product areas.

Statistical analysis integration allows product managers to move beyond simple metric comparisons to more sophisticated evaluation of experiment results. When your database analytics platform supports statistical libraries, you can implement proper significance testing, confidence intervals, and power analysis directly within your analytical workflow.

The future of self-service analytics for product teams

Database analytics represents a broader shift toward self-service analytics that empowers domain experts to answer their own questions rather than relying on centralized analytics teams. This democratization of data access is particularly valuable for product managers who need to iterate quickly on hypotheses and respond rapidly to user feedback.

The integration of AI into database analytics workflows is accelerating this trend. Natural language query interfaces, automated insight generation, and intelligent code completion are making advanced analytical capabilities accessible to product managers who may not have extensive technical backgrounds.

The transition to improved database analytics doesn't need to happen all at once, and you don't need to start from scratch with complex infrastructure setup. Quadratic provides an ideal entry point by combining the familiar spreadsheet interface with direct database connectivity and AI-powered analysis capabilities.

Product managers can begin their database analytics journey by connecting Quadratic directly to their existing data analytics database sources, such as Postgres, Snowflake, MySQL, or other systems. Rather than working through data teams to establish access permissions and learn complex SQL syntax, Quadratic's AI can help you write queries using natural language descriptions of what you want to analyze.

Start with a single high-value use case, such as analyzing a specific experiment or measuring the impact of a recent feature change. In Quadratic, you can describe your analytical needs in plain English, such as "Show me conversion rates for users who signed up last month and spent more than $1,000 during August." The AI provides both the SQL query and the resulting analysis. This approach lets you focus on interpretation and decision-making rather than query syntax.

As you become more comfortable with database analysis tools through Quadratic's interface, you can gradually expand your scope to include more complex analyses and additional data sources. The platform's combination of SQL querying, Python analysis capabilities, and AI assistance means you can grow your analytical sophistication without switching between multiple tools or learning entirely new technical stacks.

The AI-powered approach in Quadratic also ensures that your queries are optimized and follow best practices, addressing common concerns about query performance and database impact that often create barriers to direct database access. You get the benefits of expert-level database analytics without needing years of SQL experience.

The investment in learning database analytics through platforms like Quadratic pays dividends that extend far beyond individual experiments or features. Product managers who can independently access and analyze their source of truth data are better equipped to identify opportunities, validate hypotheses, and guide their teams toward impactful product decisions. In an increasingly competitive business environment where speed and precision of product learning determine success, database analytics provides a crucial competitive advantage.

How to get into data analytics: A beginner's roadmap to success

Quadratic — Mon, 28 Jul 2025 18:16:07 +0000

Have you scrolled through Netflix and realized the recommendations are actually what you want to watch? Or noticed your local coffee shop seems to magically have your exact drink ready during the morning rush? These successful predictions are the result of data analysts who work behind the scenes, and they no longer have to be computer programmers.

Consider this example. A marketing manager, who freely admits she "can barely handle Excel formulas," discovered her company's most profitable customers were not who anyone expected. She used natural language queries to an AI, such as "Which customers generate the most profit?" The AI revealed a previously unrecognized customer pattern, and that led to an increase in targeted campaign effectiveness.

This represents something fundamentally different from how to get into data analytics and the traditional career roadmap. Instead of spending months learning programming languages before you can analyze anything meaningful, modern LLM-powered data analysis platforms let you start with questions and get immediate answers. If you're wondering how hard it is to get into data analytics, the answer might surprise you: it's more accessible now than ever before.

Why learning data analytics has never been more accessible

Many people ask, "how hard is data analytics?" An accurate answer depends entirely on your approach. The traditional path was genuinely challenging with months of learning programming syntax before you could analyze your first real dataset. Today's reality tells a different story.

People who change careers to be data analysts succeed because they bring business intuition. When you understand how customers behave or how operations function, you are already thinking analytically. Understanding how to learn data analytics becomes a matter of amplifying this existing knowledge with better tools rather than starting from scratch.

Whether you are a student or changing careers, how to get into data analytics with no experience has become much more straightforward. Modern platforms bridge the experience gap by providing AI assistance that acts as both tool and mentor. You can start analyzing real data immediately while gradually building your technical knowledge.

The three foundational skills that matter most

If you are wondering how to study data analytics effectively, focus on the three must-have skills for data analysts that employers look for in interviews.

Database thinking forms the universal foundation, and it involves data preparation and data analysis. With AI tools for data analysis like Quadratic, you can ask it to tidy up the dataset. You can then ask analysis questions, such as, "Show me customers with revenue over $10,000." The AI generates the complex SQL automatically. What matters is understanding what data exists, how different tables relate to each other, and what questions will yield actionable insights. However, SQL is a core competency you need to learn.

Statistical fundamentals prevent costly analytical mistakes. For example, understand correlation versus causation, sampling bias, and statistical significance. For instance, noticing that ice cream sales correlate with drowning deaths does not mean ice cream causes drowning, although both increase during summer months. Learn statistical reasoning protects you from drawing dangerous conclusions from perfectly accurate data.

Data visualization and communication are great data analytics techniques that transform your analyses into compelling business stories. For example, learn when bar charts work better than line graphs, and why simplicity trumps complexity in executive presentations. Your goal is making the insights you discover immediately actionable for decision-makers.

How to start learning data analytics with the right resources

The abundance of learning options can feel overwhelming when you're trying to learn Python data analytics or figure out where to begin. There are resources that consistently produce job-ready analysts.

Start with Google Data Analytics Professional Certificate on Coursera for comprehensive coverage. The courses span everything from spreadsheet analysis to advanced SQL and basic programming, including hands-on projects using real datasets. This program directly answers how to get a job in data analytics by providing industry-recognized credentials.

Khan Academy's Statistics and Probability course builds essential mathematical foundations without requiring advanced calculus. This resource is perfect for anyone asking how can I improve my data literacy for analytics, as it builds the fundamental reasoning skills that underpin all analytical work.

DataCamp's "Data Analyst with Python" career track offers structured learning for those specifically interested in mastering Python for analytics. The track covers Pandas for data manipulation, NumPy for numerical operations, and Matplotlib for visualization. Their interactive coding environment lets you practice immediately rather than watching passive video lectures.

Kaggle datasets and competitions provide endless opportunities to practice. Start with beginner-friendly challenges that include detailed tutorials, then progress to more complex problems. The community forums provide invaluable exposure to different analytical approaches and professional best practices. Or, download a Kaggle dataset to practice data analysis in Quadratic.

There are many other resources available, ranging from established universities like Harvard to corporations like Microsoft to discussions on Reddit. You can learn in your own way and at your own pace while using tools like Quadratic to practice what you are learning.

How AI-powered platforms accelerate your learning journey

Here's where the game has fundamentally changed. Traditional analytics required you to master multiple disconnected tools. These included Excel for basic analysis, SQL for database queries, and Python for advanced statistics. Each transition meant learning new interfaces and debugging processes.

AI-powered platforms like Quadratic eliminate these artificial boundaries. You start with a familiar spreadsheet as shown in the below image with the AI chat interface on the left. You can use standard spreadsheet operations, and you can use the AI to seamlessly incorporate data and do statistical analyses, including charts and graphs. The AI acts as both tool and teacher, generating examples with code and explaining concepts in real-time.

Consider how this works in practice. You might begin analyzing sales data using basic spreadsheet functions, such as calculating averages and creating simple charts. When you are ready for more sophisticated analysis, you can ask the AI to generate Python code for customer segmentation. The results appear immediately in your familiar spreadsheet format, but now you are working with enterprise-grade analytical capabilities.

The AI assistance goes beyond simple code generation. It explains why certain approaches work better for specific problems, suggests improvements to your methodology, and helps debug issues that would otherwise require extensive research. This creates a learning experience that's both immediate and comprehensive.

Your first analytics project: how to get experience in data analytics

You may ask, "How to get experience in data analytics." Start with a dataset related to something you genuinely care about. This could be local housing markets, sports performance, or business metrics from your current industry. For example, an economist who was recently laid off built a model that analyzed his 10 favorite electronic dance music (EDM) songs. He then used it to search YouTube for new artists he would like.

The question of how to get into data analysis starts practically with your first real project. Begin with descriptive analysis that answers basic questions. What patterns exist in the data? How do different variables relate to each other? Build a portfolio that demonstrates your skills and be able to explain the details.

AI-powered platforms like Quadratic can accelerate your development. Traditional approaches require you to master different tools for each analytical step. With integrated AI assistance, you can perform sophisticated analysis while learning progressively. The image below shows a copy of Quadratic’s template tutorial for machine learning.

The left pane shows part of the AI’s response to the prompt “explain the chart and three steps to me.” It explained the meaning of the chart and its variables, along with the three steps necessary to create the chart and the other information shown in the right-side pane.

The AI can suggest relevant analytical approaches based on your data characteristics, generate starter code for statistical tests, and help create professional visualizations. This allows you to attempt more ambitious projects while building both technical skills and analytical judgment.

Document your entire process thoroughly. Create both technical documentation showing your analytical methodology and executive summaries highlighting key business insights. This demonstrates your ability to communicate with both technical and business stakeholders, and that is a crucial skill in professional settings.

Making the transition successful

Building a compelling portfolio requires strategic project selection. Choose analyses that demonstrate different capabilities, such as data cleaning, statistical analysis, visualization, and business insight generation. Frame projects in business terms: instead of "performed regression analysis," write "identified key factors driving customer retention."

Network actively within the analytics community through professional associations and online forums. Many positions are filled through referrals rather than public postings. Participate in discussions, share your analyses, and build relationships with established professionals.

Prepare thoroughly for interviews, which typically include both technical and behavioral components. Practice explaining your portfolio projects clearly, focusing on business impact rather than just technical details. Be ready to solve analytical problems and discuss your methodology under questioning.

Why AI represents the future of analytics

The integration of artificial intelligence into analytics platforms democratizes access to sophisticated analytical capabilities. Platforms like Quadratic exemplify this evolution. You can start with basic spreadsheet operations and progressively incorporate database queries, statistical modeling, and interactive visualizations. The AI bridges technical gaps while you focus on developing business insight and analytical judgment.

This approach particularly benefits career changers who bring strong domain expertise but limited technical backgrounds. Your ability to ask meaningful questions, understand business context, and interpret results correctly becomes more valuable than pure coding ability.

Your immediate next steps

Start today with a simple analysis of something that interests you, such as personal finances, local market trends, or data from your current role. Use tools that feel comfortable initially, then expand capabilities as confidence grows.

Focus on analytical thinking through practice rather than attempting to master every technique before starting. The field evolves rapidly, making adaptability and learning ability more valuable than encyclopedic technical knowledge.

Connect with the analytics community early. Share your work, ask questions, and learn from others' experiences. Remember that data analytics fundamentally involves solving problems and generating insights that drive better decisions.

With modern AI-powered spreadsheets removing traditional barriers, you can begin this transformation immediately while building expertise progressively. The combination of your existing knowledge, consistent practice, and AI-enhanced tools positions you to make this career transition successfully.

Quadratic AI now supports formatting

Quadratic — Tue, 27 May 2025 17:46:19 +0000

Quadratic AI now supports formatting; AI can intelligently apply much of the same formatting you otherwise would have access to. With support for:

Text styling: bold, italic, underline, strikethrough
Alignment and text wrapping
Applying color to text and cell backgrounds
Number formatting: commas, currencies, percentages, date/time formatting

Below are some scenarios that showcase the power of this feature.

Broad formatting requests

Have a sheet that could use a touch-up? Tell the AI to format your sheet however it thinks best.

Format at specific locations

Have a more opinionated location for your formatting? Tell the AI precisely where you’d like the formatting applied, and the AI will have an especially efficient time applying accurate formatting across the locations you define.

Conditional formatting

This action will usually be very accurate, but performance will degrade as you ask the AI to format larger and larger swathes of data. Expect accuracy up to hundreds or sometimes thousands of rows of data at a time.

Quadratic AI, now supporting formatting

The above are just a few possible examples of AI + formatting. Much more is possible with this feature; we can’t wait to see where else you’ll take it.

Must-have skills for data analysts in 2025: A complete guide

Quadratic — Thu, 08 May 2025 20:38:43 +0000

Do you want to stay competitive in the rapidly evolving field of data analysis? As organizations increasingly rely on data-driven decision-making, the skills needed by data analysts have become more diverse and sophisticated than ever before. In 2025, successful data analysts need a versatile combination of technical skills and soft skills to transform raw data into actionable insights that drive business value.

This comprehensive guide explores the essential data analyst skills you need to thrive in today's business environments. We'll cover both the technical foundations and the equally important soft skills that set exceptional analysts apart. This post also highlights how you can become more efficient and effective by using modern tools such as Quadratic AI.

Technical skills for data analysts

Programming and data manipulation

Technical skills for data analyst roles have become increasingly code-focused over the past few years. While spreadsheets remain valuable for quick analyses, programming languages offer greater flexibility, reproducibility, and the ability to handle larger datasets.

Programming languages are the foundation of data analytics skills. SQL is the primary language for querying databases, and it enables analysts to extract the data they need from various sources. A strong command of SQL enables you to write complex queries with joins across multiple tables, perform aggregations and calculations, create and modify database structures, and optimize queries for better performance.

Python is the most versatile language for data analysis. Being a data analyst with Python and SQL skills makes you especially valuable to employers. Key Python libraries that every analyst should know include Pandas for data manipulation and analysis, NumPy for numerical calculations, Matplotlib/Seaborn for data visualization, and Scikit-learn for implementing machine learning algorithms.

The R language is useful for statistical analysis and data visualization. Many organizations use both languages, so familiarity with R can be an advantage in certain roles.

Data visualization

The ability to create compelling visualizations is among the top skills for data analyst positions. Effective data visualization transforms complex findings into intuitive visual stories that stakeholders can understand and act upon. Key data visualization software include Tableau and Power BI for creating interactive dashboards, as well as Python libraries like Matplotlib, Seaborn, and Plotly.

Modern Power BI and Tableau alternatives like Quadratic combine the best of spreadsheets and programming, allowing analysts to create different types of visualizations directly from their data using code or AI assistance, making it easier to communicate findings effectively.

Statistical analysis and machine learning

A strong foundation in statistics is essential for meaningful data analysis. Understanding concepts like descriptive statistics, probability distributions, hypothesis testing, regression analysis, and experimental design helps you avoid common pitfalls in data interpretation and ensures your analyses lead to valid conclusions.

While deep expertise in data analysis doesn't necessarily require advanced machine learning skills, understanding the fundamentals is increasingly important. Familiarity with supervised and unsupervised learning approaches, common algorithms like linear regression and decision trees, feature selection and engineering, model evaluation metrics, and basic neural networks enables you to move beyond descriptive analytics to predictive and prescriptive approaches that provide greater business value.

Data cleaning and preparation

Data preparation often consumes a large portion of an analyst's time. Proficiency in cleaning, transforming, and structuring data is a critical skill set. This includes handling missing values, detecting and addressing outliers, normalizing and standardizing data, feature engineering, and managing data types and formats.

Data cleaning tools like Quadratic streamline these processes through AI assistance, allowing analysts to focus more on generating insights rather than mundane preparation tasks.

Data governance and security awareness

In today's regulatory environment, understanding data governance principles has become an essential component of the data analyst’s skill set. Analysts must be familiar with regulations like GDPR (General Data Protection Regulation), CCPA (California Consumer Privacy Act), and industry-specific compliance requirements that govern data usage.

This includes implementing appropriate data anonymization techniques, maintaining proper documentation of data lineage, understanding consent requirements for data usage, and collaborating with legal and security teams to ensure compliance. As organizations face increasing scrutiny over their data practices, analysts who can navigate these complexities while still delivering insights become particularly valuable.

Cloud computing environments

Modern data analysis increasingly takes place in cloud environments. Familiarity with major cloud platforms like AWS, Azure, and Google Cloud Platform has become one of the key technical skills for data analyst roles. This includes working with cloud-based data warehouses, understanding serverless computing options, configuring and optimizing cloud resources, and leveraging cloud-native analytics services.

Advanced SQL capabilities

While basic SQL is mentioned earlier, advanced SQL capabilities separate junior analysts from experienced professionals. These advanced skills include writing complex window functions for sophisticated calculations, using Common Table Expressions (CTEs) for more readable and maintainable queries, optimizing query performance for large datasets, understanding database indexing strategies, and leveraging specialized functions for different systems. These capabilities are essential skills required for data analyst positions working with enterprise-scale data.

Soft skills for data analysts

While technical capabilities form the foundation of your data analyst skill set, soft skills for data analyst roles often determine your ultimate impact and career trajectory.

Communication and business acumen

Being a data analyst with strong communication skills sets you apart in the job market. The ability to translate complex findings into clear, actionable insights is invaluable. This means crafting compelling data stories, adapting your communication style to different audiences, creating clear documentation, presenting findings confidently, and explaining technical concepts in accessible language.

Understanding the business context transforms a technically competent analyst into a strategic partner. Strong business acumen includes knowledge of industry trends and challenges, understanding of key performance indicators, ability to connect analytical findings to business objectives, awareness of stakeholder priorities, and insight into how decisions impact different areas of the organization.

Critical thinking and project management

Data analysis is fundamentally about solving problems. Strong critical thinking helps you formulate the right questions to guide your analysis, identify patterns and anomalies in data, question assumptions and recognize biases, draw meaningful conclusions, and propose actionable recommendations.

Data analysis projects often involve multiple stakeholders, datasets, and timelines. Effective project management means scoping projects effectively, breaking down complex analyses into manageable tasks, prioritizing work based on business impact, managing timelines and expectations, and collaborating effectively with cross-functional teams.

Adaptability and continuous learning

The field of data analysis evolves rapidly. Successful analysts demonstrate a willingness to learn new tools and technologies, flexibility in approaching different types of problems, openness to feedback and new perspectives, resilience when facing challenges, and commitment to staying current with industry developments.

Data storytelling

Beyond basic communication, data storytelling has emerged as a critical soft skill for data analysts. This involves structuring analyses as compelling narratives with clear beginnings, middles, and ends; selecting visualizations that best support your key points; creating a logical flow of information that guides audiences to important conclusions; incorporating relevant context and business implications; and using appropriate analogies and examples to make complex concepts relatable. Skilled data storytellers can transform even the most technical analyses into memorable presentations that drive action.

Essential tools for modern data analysts

The skill set for data analyst roles encompasses proficiency with various tools that enhance productivity and capabilities. These include SQL databases like MySQL and PostgreSQL, cloud data warehouses such as Snowflake and BigQuery, and visualization platforms like Tableau and Power BI.

Version control and collaborative tools

As data analysis becomes increasingly collaborative, proficiency with version control systems like Git has become an important part of the data analyst's technical skills toolkit. Understanding how to manage code repositories, track changes to analyses, collaborate on shared codebases, resolve conflicts, and maintain documentation helps ensure reproducibility and facilitates teamwork. Beyond Git, familiarity with project management tools, documentation platforms, and collaborative environments enables analysts to work effectively within cross-functional teams.

Data engineering fundamentals

While dedicated data engineers typically build and maintain data infrastructure, successful analysts benefit from understanding the fundamentals. This includes knowledge of ETL/ELT processes, data pipeline architecture, basic data modeling concepts, awareness of database optimization principles, and familiarity with data orchestration tools. These skills help analysts work more effectively with engineering teams and sometimes implement simpler data pipelines themselves, expanding their skills for data analyst roles into more technical territories.

Modern integrated platforms like Quadratic are changing how analysts work by combining spreadsheet functionality for familiar workflows, code execution capabilities for Python, SQL, and JavaScript, AI assistance for data preparation and analysis, visualization tools for creating compelling charts, and collaboration features for team-based analysis. These integrated environments help streamline the analysis process and make advanced data analyst technical skills more accessible, even to those early in their career.

How to improve data analysis skills

Looking to enhance your capabilities? There are several effective ways to learn how to improve data analysis skills. Structured learning through online courses on platforms like Coursera or DataCamp, relevant certifications in SQL or Tableau, bootcamps focused on data analytics, or advanced degrees for specialized roles provide a solid foundation.

Practical application is equally important. Working on personal projects using real datasets, participating in data science competitions on platforms like Kaggle, contributing to open-source data projects, and creating a portfolio showcasing your work helps solidify theoretical knowledge through hands-on experience.

Community engagement accelerates growth through joining data science communities in Slack groups or Reddit forums, attending meetups and conferences, participating in hackathons, and following industry leaders. Modern tools like Quadratic can enhance your learning by providing an integrated environment for various languages, offering AI assistance for code generation and analysis, supporting collaboration with peers, and streamlining workflows to focus on analytical concepts rather than tool-specific complexities.

Skills for your data analyst resume

When crafting your resume, highlighting the right skills for a data analyst resume is crucial. Consider organizing your data analyst skills bullet points into clear sections that showcase both technical and soft skills.

For technical data analytics skills, focus on programming languages like SQL and Python, data visualization tools such as Tableau and Power BI, statistical analysis and modeling techniques, database management and data warehousing experience, and proficiency with platforms like Quadratic and Excel. For soft skills, emphasize communication abilities including data storytelling and presentation, problem-solving and critical thinking capabilities, project management experience, collaboration and teamwork, and your adaptability and commitment to continuous learning.

In your projects section, document experiences that demonstrate your skills in action by describing the problem addressed, methods and tools used, key findings and impact, and quantifiable results where possible. Remember that the skills required for data analyst positions vary across industries and organizations, so tailor your resume to highlight the most relevant skills for each role you apply for.

The future of data analysis skills

The landscape of skills needed for data analyst roles continues to evolve. Several trends will shape the field in 2025 and beyond.

AI integration is increasingly transforming the data analysis workflow. Rather than replacing analysts, AI tools augment human capabilities by automating routine data preparation tasks, suggesting visualizations and analyses, generating code and SQL queries, identifying patterns and anomalies, and enhancing data exploration. Platforms like Quadratic exemplify this trend, providing AI assistance that helps analysts focus on higher-value tasks while handling routine operations more efficiently.

As data usage expands, skills related to ethical data practices and governance are becoming essential. Understanding privacy regulations, implementing data protection measures, recognizing and addressing algorithmic bias, ensuring transparent data practices, and maintaining data quality and lineage are increasingly critical components of an analyst's skill set.

While foundational skills remain important, many analysts are developing expertise in specific domains such as marketing analytics, financial analysis, healthcare data analysis, supply chain analytics, or people analytics. Domain specialization combines technical skills with industry-specific knowledge, creating unique value propositions for analysts.

Experimental design and A/B testing

As organizations seek to make more data-driven decisions, rigorous testing methodologies have become increasingly important skills needed for data analyst positions. Proficiency in experimental design includes formulating testable hypotheses, determining appropriate sample sizes, designing controlled experiments, implementing randomization techniques, and analyzing results with statistical rigor. A/B testing specifically has become essential for product, marketing, and UX teams, making analysts who can design, execute, and interpret these tests particularly valuable in driving evidence-based decisions.

Industry-specific analytical tools

While core data analytical skills remain consistent across sectors, the rising importance of industry-specific tools is creating new specialization opportunities. Depending on your focus area, this might include Google Analytics and attribution modeling tools for marketing analysts, Bloomberg Terminal and financial modeling packages for financial analysts, healthcare-specific data systems for medical analysts, or IoT analytics platforms for manufacturing analysts. These specialized toolsets combine with domain knowledge to create unique expertise that can significantly enhance your data analyst skills resume for targeted positions.

Conclusion

The most successful data analysts in 2025 will combine strong technical foundations with exceptional soft skills and domain knowledge. By developing both data analyst technical skills and data analyst soft skills, you'll be well-positioned to thrive in this dynamic field.

Modern tools like Quadratic are democratizing access to advanced analytics capabilities, allowing analysts to focus more on generating insights and less on technical complexities. By leveraging these tools alongside continuous learning, you can stay ahead of the curve in your data analysis career.

Remember that becoming an exceptional analyst is a journey of continuous improvement. Focus on building a balanced skill set for data analyst roles by combining technical expertise with communication skills, business acumen, and analytical thinking. This holistic approach will make you invaluable to any organization that values data-driven decision-making.

Ready to enhance your data analysis capabilities? Try Quadratic today and experience how the right tools can accelerate your analytical workflow and help you focus on what matters most: transforming data into actionable insights that drive business value.

The ultimate guide to AI spreadsheet analysis

Quadratic — Mon, 14 Apr 2025 20:05:52 +0000

In today's data-driven business environment, the ability to extract actionable insights from your data has become a competitive necessity. Analysts and business professionals often spend countless hours wrestling with complex datasets, struggling through multiple tools and processes before uncovering insights with business value.

This is where AI spreadsheet analysis is changing the game. By integrating artificial intelligence directly into the familiar spreadsheet interface, platforms like Quadratic AI are dramatically simplifying every phase of the data workflow, from initial preparation to final visualization and decision support.

Understanding modern AI spreadsheet analysis

Traditional spreadsheets have served as the cornerstone of business analysis for decades, but they have significant limitations. Complex analyses require specialized knowledge, large datasets cause performance issues, and collaboration often leads to version-control nightmares. Modern AI spreadsheet analyzers preserve the intuitive grid interface while leveraging AI to overcome these constraints and add a natural language interface. The following screenshot of calculating the ROI for different platforms is from a Quadratic AI spreadsheet.

Quadratic exemplifies this new generation of LLM-powered analysis tools. By combining a familiar spreadsheet interface with powerful AI capabilities, Quadratic creates an environment where technical and non-technical users can perform sophisticated analyses with unprecedented ease. The platform's AI actively assists in every phase of your analytical workflow by suggesting approaches, identifying patterns, and translating natural language requests into powerful analyses.

This integration of AI for spreadsheet analysis throughout the entire data analytics lifecycle transforms how organizations approach analytics by democratizing access to advanced capabilities while accelerating time to insight. Let's explore how AI that analyzes spreadsheets simplifies each critical phase of your data workflow.

Data collection and integration

Modern analytics rarely relies on a single data source. Organizations typically need to combine information from multiple systems, databases, and external sources to create a comprehensive view of their operations. Traditionally, this has required complex ETL (Extract, Transform, Load) or ELT (Extract, Load, Transform) processes, often involving specialized tools and technical expertise. This lack of data accessibility often created bottlenecks in the analytical workflow.

AI tools for data analysis like Quadratic streamline this process through intelligent connectivity. Direct database connections allow you to seamlessly integrate data from PostgreSQL, MySQL, Snowflake, and other sources without leaving your spreadsheet environment. Rather than writing complex SQL queries, you can use natural language to specify the data you need, and the AI model will suggest ideas. The following screenshot shows a combined dataset from three sources with email as the common id. It also shows the chat interface with the AI’s suggestions.

The true power of this approach becomes apparent when, for example, a marketing analyst blends CRM data with website analytics and social media metrics. With traditional tools, this could involve exporting multiple files, creating lookup tables, and performing manual joins. With AI for data analysis, the process becomes conversational: "Combine our customer data with website visit metrics and social engagement, matching on customer email address."

Quadratic's implementation of this capability creates a seamless experience. The platform handles the technical complexity of establishing connections, matching records across sources, and formatting the combined dataset for analysis with SQL queries. This accelerates the integration process and maintains live connections, ensuring your analysis always reflects the most current information.

Data cleaning and preparation

Data preparation has traditionally been the most time-consuming aspect of any analysis project. Raw data typically arrives with missing values, inconsistent formatting, duplicate entries, and outliers that must be identified and addressed before analysis can begin. This tedious process has long been the necessary but dreaded first step in any analytical process.

Using an AI to analyze spreadsheet data fundamentally transforms this experience. Data cleaning tools like Quadratic can automatically scan your dataset upon import, identifying quality issues and suggesting remediation strategies. Rather than manually hunting for problems, you ask the AI to "clean this dataset" or "standardize these date formats." It executes these operations in seconds rather than hours, such as standardizing the date formats shown in the previous image.

Consider the experience of analyzing customer-purchase data. In a traditional spreadsheet, you might spend hours standardizing product categories, converting date formats, and handling missing values. With an AI analyzing data, you upload your raw data and ask: "Clean this dataset by standardizing date formats, filling missing values with appropriate measures, and identifying outliers in the purchase amounts." The AI quickly executes these operations and provides clear documentation of the changes made. This allows you to review and adjust as needed.

This AI-powered data analysis approach to data preparation saves time and improves quality by applying consistent methodologies and highlighting issues that might be missed. Analysts can spend more time on actual analyses, where their expertise adds more value.

Exploratory data analysis

Effective exploratory analysis requires creativity and rigor, which refer to the carefulness, consistency, and transparency to ensure the accuracy and unbiased interpretation of the data. It requires the ability to ask insightful questions while systematically examining relationships within the data.

Traditional spreadsheets offer basic tools for this exploration, but discovering meaningful patterns often requires expertise in statistical methods and visualization techniques that many business users lack. Users may understand the mathematics without having the technical skills to do them within a spreadsheet.

Using an AI transforms exploratory analysis from a specialized skill to an intuitive conversation. Instead of manually creating pivot tables or writing complex formulas, users can directly ask questions of their data: "What factors are most strongly correlated with customer churn?" or "Show me the unusual patterns in our website traffic over the past month."

The AI can provide the requested information, and it also can guide the exploration process by suggesting additional questions to investigate. For example, when examining sales data, the Quadratic AI might automatically identify and suggest seasonality patterns, unusual customer segments, or emerging trends that need to be explored.

This guided data exploration is particularly valuable for domain experts who understand their business but may lack statistical expertise. A product manager analyzing spreadsheet data with AI can explore whether the data has meaningful patterns without having the skills to calculate correlation coefficients or clustering algorithms. The AI handles the analytical complexity, allowing the manager to focus on interpreting the business implications of the findings.

Advanced analysis and modeling

Statistical analysis forms the foundation of data-driven decision-making, but terms like regression analysis, hypothesis testing, and variance analysis can intimidate non-specialists. Implementing these techniques correctly requires specialized knowledge that many business users lack.

However, with natural language tools such as Quadratic AI, non-specialists can gain the benefits from data analytics. A Harvard Business School webpage offers a free downloadable ebook that teaches data literacy and suggests further resources. It defines four key categories of data analytics:

Descriptive: What happened?
Diagnostic: Why did this happen?
Predictive: What might happen in the future?
Prescriptive: What should we do?

These are important to learn if you do not already have data analytic skills. The report cites a study in which "56 percent of respondents said data analytics led to 'faster, more effective decision-making’ at their companies. Other benefits cited include:

Improved efficiency and productivity (64 percent)
Better financial performance (51 percent)
Identification and creation of new product and service revenue (46 percent)
Improved customer acquisition and retention (46 percent)
Improved customer experiences (44 percent)
Competitive advantage (43 percent)"

AI tools for analyzing data allow you to do standard and advanced statistical capabilities because they translate natural language requests into mathematical operations. You can simply describe what you want to understand: "Is there a significant difference in conversion rates between our two landing page designs?" or "What factors best predict customer lifetime value based on our historical data?"

You can also ask it to suggest what statistics it thinks should be calculated on the dataset. Even advanced experts can benefit from this when the AI knows there are patterns in the data that the user has not been shown. The AI can suggest an analysis the expert had not considered doing.

Quadratic's implementation of this capability goes beyond this. The AI data analyzer provides appropriate statistical methods based on the nature of your data and question, ensuring methodological validity. It also explains the results in business terms rather than statistical jargon. For example, when comparing the performance of two marketing campaigns, the platform might automatically conduct a t-test, explain the significance of the results, and highlight practical implications for future marketing investments.

This approach extends to more sophisticated analytical techniques as well. Regression analysis, typically requiring specialized statistical software, becomes accessible through natural language: "Build a model that predicts monthly sales based on marketing spend, seasonality, and economic indicators." The AI handles the technical implementation by selecting the appropriate variables, transforming data as needed, and validating the model. Again, it presents its results in an accessible format that focuses on business insights rather than statistical mechanics.

Predictive analytics

The ability to forecast future outcomes based on historical data represents one of the most valuable applications of analytics. Traditional approaches to predictive modeling typically require specialized expertise in statistical methods or machine learning, creating a significant barrier for many organizations.

AI spreadsheet analyzer platforms like Quadratic bring predictive capabilities directly into the familiar spreadsheet environment, making forecasting accessible to a much wider audience. Users can leverage sophisticated predictive algorithms without needing to understand their internal workings, focusing instead on the business questions they need to answer: "Forecast our expected sales for the next six months based on historical patterns" or "Predict inventory requirements for our upcoming product launch."

The implementation of these capabilities combines power with accessibility. Behind the scenes, the AI that can analyze spreadsheets might employ time series analysis, machine learning algorithms, or statistical models, automatically selecting the most appropriate approach based on your data characteristics and the specific forecasting task. The results are presented in business-friendly formats, often as clear visualizations with confidence intervals and explanatory text that highlights key factors influencing the predictions.

This democratization of predictive capabilities enables organizations to incorporate forward-looking insights into their regular planning processes. A retail manager can generate seasonal sales forecasts without waiting for the analytics team. A supply chain planner can predict potential inventory shortages weeks in advance. By embedding these capabilities directly in the spreadsheet environment, predictive analytics becomes an everyday tool rather than a specialized project.

Text and qualitative analysis

Unstructured text data represents a vast source of potential insights for many organizations. Customer feedback, support tickets, survey responses, and social media comments all contain valuable information that can guide product development, marketing strategies, and customer service improvements. However, traditional spreadsheets offer minimal support for analyzing text data beyond basic keyword searches.

AI for qualitative data analysis platforms like Quadratic provides sophisticated natural language processing capabilities. Users can analyze text fields as easily as numerical data, uncovering patterns, sentiments, and themes without specialized text analytics expertise. Rather than manually reading thousands of comments, you can ask the AI chatbot for data analysis to "Identify the main themes in our customer feedback" or "Analyze the sentiment of these product reviews and show how it varies by product category."

The following images from Quadratic AI show a set of feedback comments from customers, a histogram chart comparing the number of neutral, positive, and negative comments, and the specific feedback related to the product’s features.

The implementation of these capabilities transforms how organizations leverage qualitative data. Sentiment analysis can automatically classify text as positive, negative, or neutral, revealing trends in customer satisfaction across products or time periods. Topic modeling can identify key themes and concerns, highlighting areas that require attention. Entity extraction can recognize mentions of specific products, features, or competitors, creating structured data from unstructured text.

This ability to systematically analyze a spreadsheet for qualitative data alongside quantitative metrics creates a more comprehensive view of business performance. For example, a product team can know how many customers are using a feature and how they feel about it. A marketing team can identify specific product attributes that drive positive sentiment. By integrating text analysis directly into the spreadsheet environment, these qualitative insights become accessible throughout the organization rather than remaining siloed in specialized text analytics tools.

Visualization with communication

Effective data visualization transforms raw numbers into compelling visual narratives that highlight patterns, relationships, and trends. However, creating impactful visualizations traditionally requires both design expertise and technical skills. Trained data analysts would know which chart types best represent different data relationships, how to configure axes and scales appropriately, and how to highlight key insights without creating visual clutter.

AI qualitative data analysis platforms like Quadratic simplify this process through intelligent visualization recommendations and natural language generation. Rather than manually configuring chart properties, users can simply describe what they want to see: "Create a visualization showing sales trends by region over the past year" or "Show me the relationship between marketing spend and customer acquisition cost."

The implementation of these capabilities combines analytical rigor with visual effectiveness. For example, an AI tool to analyze excel data can recommend appropriate visualization types, such as a line chart for time series data, a scatter plot for examining correlations, or a stacked bar chart for comparing composition across categories.

This approach makes sophisticated visualization accessible to all users, regardless of their technical background. A finance team can quickly create compelling charts for board presentations. A marketing manager can visualize campaign performance across multiple dimensions. By reducing the technical barriers to effective visualization, AI spreadsheet analysis ensures that insights are communicated clearly and persuasively throughout the organization.

Collaboration and knowledge sharing

Effective analysis rarely happens in isolation. Teams need to collaborate on data exploration, share insights, and build collective understanding to drive informed decisions. Traditional spreadsheets create significant friction in this collaborative process because files get emailed back and forth, changes are difficult to track, and knowledge remains siloed in individual workbooks.

AI for spreadsheet analysis platforms like Quadratic transform collaboration through real-time multiplayer capabilities, intelligent insight sharing, and knowledge capture. Team members can work simultaneously on the same analysis, seeing each other's changes in real-time and building on each other's ideas. Rather than emailing static reports, they can share live dashboards that update automatically as new data becomes available.

The implementation of these capabilities creates a truly collaborative analytical environment. Comments and questions can be attached directly to specific cells or visualizations, creating context-rich discussions that enhance understanding. Version history tracks changes over time, allowing teams to understand how analyses have evolved and revert to previous versions if needed. AI-powered suggestions can highlight insights that might be relevant to specific team members based on their roles or previous interests.

This collaborative approach accelerates the journey from data to decisions by bringing diverse perspectives together around a shared analytical canvas. A product team can collectively explore customer feedback, building a richer understanding of user needs. A cross-functional task force can develop and refine forecasts, incorporating insights from multiple departments.

The future of AI-powered data analysis

The integration of artificial intelligence into spreadsheet analysis represents more than just an incremental improvement in analytical tools. It signals a fundamental shift in how organizations approach data. By simplifying complex analytical tasks, automating routine processes, and making advanced capabilities accessible to users of all technical levels, platforms like Quadratic are democratizing data analysis and accelerating the journey from information to insight.

This transformation carries profound implications for organizations seeking to become truly data-driven. Analysis is no longer constrained by technical bottlenecks or specialized expertise. Anyone with domain knowledge can ask sophisticated questions of their data and receive immediate, actionable insights. Analytical workflows become more efficient, freeing time for deeper exploration and strategic thinking. Collaboration becomes seamless, ensuring that insights are shared, refined, and applied throughout the organization.

The future of AI spreadsheet analysis will bring even greater capabilities. These will include more sophisticated predictive models, deeper integration with operational systems, and increasingly conversational interfaces that make complex analysis feel as natural as asking a question. But the fundamental value proposition will remain the same: simplifying the entire data workflow to help organizations transform raw information into meaningful business impact.

For organizations seeking to enhance their analytical capabilities and accelerate data-driven decision-making, free AI tools for data analysis like Quadratic offer a compelling solution. By embracing these advanced tools, organizations can ensure that their data becomes a strategic asset rather than an untapped resource, driving innovation, efficiency, and competitive advantage in an increasingly data-driven world. More team members can be empowered, and that empowerment fuels better decisions across the entire organization.

Cursed Excel: “1/2”+1=45660

Quadratic — Tue, 08 Apr 2025 20:42:10 +0000

Quadratic aspires to be the best spreadsheet for data analysis, which implies two conflicting goals:

Maintain feature parity with Microsoft Excel
Be good

I mean no offense to the authors of Excel; it is fantastic software that we at Quadratic very often hail as the gold standard of spreadsheet interaction. That said, it's reaching the ripe old age of 40 this year and its semantics seriously suffer from a decades-long accumulation of backwards-compatible cludges.

One of my favorite things about working here is that I get to reverse-engineer Excel, find strange quirks in its behavior, and decide what to do about them in Quadratic. I suffer every day so that our users may live blissfully unaware of the undocumented sins committed by Microsoft in the name of compatibility. Today you will gain a glimpse into the horrors I contend with, and then you too will live in fear of Microsoft Excel — not because you lack knowledge, but because you know too much.

Magic numbers

For many years, geneticists have struggled with Excel's overeager date parsing applying to names like MARCH1 or SEPT2 that aren't meant to be dates. But Excel's date parser has much weirder edge cases.

If we type ="1/2" into a cell, then of course it contains the text "1/2".

What if we add 1 to that?

45660? What?? Here's a hint: if you try this in the future, you may get a different number.

And it's not just dates! Sometimes Excel's time parser bites off a little more than it can chew. Of course typing 10:25 into a cell results in the time 10:25 a.m., but what happens if we type 10:75?

0.46875?? Where the heck did that come from?

I promise that I will explain what's going on here, but first we need to cover some technical documentation and some Catholic Church history.

(Don't) read the manual

In both of these scenarios, we're tricking Excel into parsing our input as a date or time, but displaying it as a number. As the official documentation for the DATEVALUE() function explains:

Microsoft Excel stores dates as sequential serial numbers so they can be used in calculations. By default, December 31, 1899 is serial number 1, and January 1, 2008 is serial number 39448 because it is 39,448 days after January 1, 1900.

This is helpful but contains two inaccuracies. The first is that serial number 1 represents January 1, 1900, not December 31, 1899. In fact, Excel will never display a date before 1900 and instead insists that serial number 0 represents January 0, 1900.

Thankfully, the mistake about serial number 1 is corrected elsewhere in the documentation for MONTH() and many other functions, but there is another inaccuracy that is still present, and it is much more insidious.

There are actually only 39,446 days between January 1, 1900 and January 1, 2008, not 39,448. I can understand an off-by-one error, but why is Excel off by 2?

Imagine I'm assigning a number to each day of the week. Monday is 1, Tuesday is 2, ..., and Friday is 5. But you wouldn't say that Friday is 5 days after Monday. It's only 4 days after, and you can calculate that by subtracting Monday's number from Friday's number: 5 - 1 = 4. The same thing is happening here: to get the number of days between January 1, 1900 and January 1, 2008, we should really subtract the 1900 number from the 2008 number: 39448 - 1 = 39447. This is closer, but still off by one. To understand the remaining off-by-one error, we need to grab our time machine and travel almost 450 years into the past.

Calendar systems

In October 1582, Pope Gregory XIII officially decreed that the Catholic Church would use the new calendar system developed by Aloysius Lilius (then deceased) and Christopher Clavius. The Julian calendar, which had a leap year every 4 years, had been in use for more than 1600 years but had caused so much drift that Easter had fallen out of alignment with the Spring equinox. The newly christened Gregorian calendar corrected the drift by adding a new rule: every year divisible by 100 is not a leap year, except years divisible by 400 which are still leap years. This is why the year 2000 was a leap year (because it is divisible by 400), but 1900 was not.

In 1983, almost exactly 400 years after the new calendar was adopted, Lotus Software released Lotus 1-2-3, a revolutionary spreadsheet + database + charting program. Unfortunately, news of the 1582 promulgation had not yet reached the developers of Lotus 1-2-3, so they assumed that 1900 (being a multiple of 4) was a leap year. A few years later, Microsoft released the first version of Excel with the same mistaken leap year. If you enter Feb 28, 1900 into Excel and add one, you'll get Feb 29, 1900 — a day that never happened, but is necessary to maintain compatibility with Lotus 1-2-3.

This explains the other off-by-one error in the Excel documentation; Excel is counting an extra day in February 1900, so serial numbers for dates later than that are 1 larger than you'd otherwise expect.

What happened?

Back when you still had your sanity, I promised to explain why "1/2"+1 equals 45660 and why Excel turned 10:75 into 0.46875.

The first one should actually make some sense once you realize that Excel is parsing 1/2 as January 2, 2025 (the year that I'm writing this). When we add 1 we get January 3, 2025, and there were 45,658 days between that day and January 1, 1900. Add 2 for the reasons described earlier and we get 45,660, exactly what Excel says. I don't know why it displays a number instead of a date, but the number at least makes sense.

The second one requires a deep philosophical insight: what is a time, other than a fraction of a day? For example, 6:00 a.m. is 0.25 days, so Excel represents it using the number 0.25. By this logic, 0.46875 should represent 11:15 a.m., which is 75 minutes after 10:00 a.m. so that's sort of like 10:75 a.m. if you don't think too hard about it. But deep down inside, Excel knows that this is very wrong so it displays it as a number instead.

We can even get times beyond 11:59 p.m. by using hours greater than 23. Typing 37:30 into a cell produces the number 1.5625, which represents 1:30 p.m. the next day. The number 1 represents exactly midnight at the beginning of January 1, 1900, so 1.5625 represents 1:30 p.m. on January 1, 1900.

What about Google Sheets?

Google Sheets had the brilliant idea to remove February 29, 1900 by shifting the first two months of 1900 over by one, so it represents January 1, 1900 using serial number 2 instead of 1. This is a pretty clever solution, although it's a bit awkward to start at 2 and it causes dates before March 1, 1900 to be off-by-one when importing from Excel.

What about Quadratic?

We're building Quadratic from the ground up to work well with Python, SQL, JavaScript, and other modern programming and database tooling, so incorrect calendar systems are not an option. We use the battle-tested chrono library for datetimes in Rust, which plays nicely with Python's built-in datetime library and similar data types in other languages. When importing files from Excel, we convert any cells with date formatting into the corresponding datetime. In an effort to restore balance to the universe, February 29, 1900 is converted to February 28, 1900.

Using a proper datetime system has the added bonus of letting us represent dates much farther in the past than 1900, although I'd be careful with anything before 1582. Building a spreadsheet from scratch is challenging and takes a long time to get right, so if you have a use case we don't support yet, let us know on our community forums or submit a code contribution on GitHub!

Demystifying AI model selection

Quadratic — Tue, 01 Apr 2025 21:20:37 +0000

AI model selection can be a daunting task. With so many models, so many versions of the same model, and so many benchmarks, how do you decide what to pick? And what are these obnoxiously long names like Qwen2.5-VL-32B-Instruct-bnb-4bit?

HuggingFace is the best place to explore various models that are available for fine-tuning and further study. It can be a daunting site before you understand what these model specifics mean.

Fortunately, there’s (usually) a method to the madness. Below, we outline how to think about AI model selection for any given use case.

This article is part of a series on fine-tuning tips and best practices. See the most recent article in the series where we discussed AI dataset construction here.

1. Eval vibe checks

The first step to selecting an AI model is performing a vibe check on the state of the art (SOTA). Checking the latest model releases from reputable AI players is usually a safe start. Evals aren’t everything, but they are worth looking at as a starting point. Often, companies will share their evals as big charts, usually quite favorably to themselves, leaving out competitor models that outperform them and evaluations that they underperform on. You’ll need to dig and build your comparisons as you pit models against one another.

Here’s a neutral eval that showcases the capabilities across the entire Qwen2.5 Coder line - from the smallest parameter version 0.5B to 32B.

There are many evals out there, but many are quite common. Here are a few of the most common code evals briefly explained:

HumanEval: Small Python-based coding eval by OpenAI
MBPP: Basic Python code but larger than HumanEval by Google
EvalPlus: Larger Python eval meant to catch models that are overfitted to HumanEval and MBPP by OpenCompass
MultiPL-E: HumanEval translated to a bunch of coding languages by HuggingFace and BigCode
McEval: Multiple choice (Mc) eval to test reasoning, tests a model’s ability to pick the best code, not generation by OpenCompass
LiveCodeBench: Multi-language code eval with lots of cases by Tsinghua University
CodeArena: Crowd-sourced human evaluation is where humans rank the outputs of two models and pick the best solution.

The primary goal of looking at evals is finding a few model families you’d like to experiment with and then, within those families, seeing the performance gaps between models of various sizes within the same model family. It’s good to understand the basics of evals to understand which ones ensure a model isn’t overfitted to the most popular benchmarks like HumanEval.

2. Model size

It’s generally the case that the fewer parameters a model has, the smaller its size, and thus, the lower its memory footprint and higher its speed. If your problem is easily solved with a 7B model, then it’d make no sense to run a 32B model; that’d be a waste of money and would slow down responses.

Many of the advantages of fine-tuning come from achieving the speed, cost, and memory footprint of a much larger model out of a smaller model. General advice is to start small and work your way up to larger models as needed.

It does seem that the ability to solve complex problems starts to emerge around ~10B parameters, but increasingly, models are being released that are smaller and smaller with more capabilities. Play around with various model sizes and see where you land for your use case. Often, model size is a much more relevant consideration than model family, especially for models released at similar times.

3. Base vs Instruct vs Chat vs Code

First, some definitions:

Pre-training is the main step that model providers take for training. It is done on huge data corpora, usually takes a long time, and requires huge amounts of compute.
Fine-tuning: task-specific training that model providers can do or, more commonly, by developers downstream

Models come in a few different forms - some are just the Base model as it was initially pre-trained, and others are fine-tuned versions from the model provider or from companies and hobbyists with specific use cases.

These are the most common suffixes in a model ID related to how a model was trained.

Base: This is the pre-trained model, usually without fine-tuning, trained on large data corpora like literature, code repos, social media, etc.
Instruct: This is a base model that has been fine-tuned for instruction following, e.g. assistants
Chat: self-explanatory, base model fine-tuned for chat applications (multi-turn chats)
Code: self-explanatory, base model fine-tuned to focus on coding
VL: vision language, meaning the model is multi-modal

It often makes sense for fine-tuning applications to start with Instruct models, especially when your dataset (we talked about constructing AI training datasets here) is smaller and you want to improve the instructions to work for your use case.

4. Model purpose & training data

Model providers often focus on specific use cases to build models that are especially capable of certain tasks. For example, Claude is well known to outperform GPT-4o in most coding tasks, but OpenAI has plenty of writing scenarios where it’s deemed superior. The same can be seen in open source models; some examples:

Alibaba’s QwenCoder: as the name suggests, this model excels at code tasks
Google Gemma: advanced multi-modal capabilities
Microsoft Phi-4: for code, the training data is heavily biased toward Python

Many of the most popular benchmarks are focused on Python, so it’s not surprising that the model providers are especially attentive to performing well with Python.

5. Quantization

Quantization is an option if we need to reduce our model footprint. Doing so can significantly reduce the model size, with a minor negative impact on model performance but a large positive impact on speed.

Float16 is a very standard high-bit data format. You can expect to use float16 as a standard for commercial applications. But sometimes, you might not have the hardware available to host at float16, so you can quantize down to smaller formats as necessary (while being okay with the minor performance hit). These quantized models will run significantly faster and require much less space. Some standard formats to quantize into include int8 and 4-bit. If you see a model name with a suffix like 4bit at the end, it’s a quantized version of the model to 4-bit.

6. Licenses

Author of this post is not a lawyer. Take license opinions with a grain of salt. Consult with your own counsel.

Licenses are an important consideration depending on your use case. They can span everywhere from fully permissive to completely non-permissive. Below are some examples to showcase why licenses might impact the model you choose among the current SOTA.

MIT (open source, fully permissive): Some of ****Microsoft’s Phi models
Apache 2.0 (open source, very permissive): (Most of) Alibaba’s Qwen models, Mistral
Custom licenses (sometimes very permissive, other times not): Google’s Gemma, Meta’s Llama
Fully proprietary: Vast majority of OpenAI and Anthropic’s models

Depending on your use case, whether you’re a hobbyist or a company, and many other considerations, you’ll want to find the license that best fits your use case.

Not your weights, not your model.

Conclusion

Hopefully, you’re walking away with some ideas on how to improve your model selection. Ideally, you’ve walked away from this article understanding how to read a model ID like Qwen2.5-VL-32B-Instruct-bnb-4bit. Here’s the reminder for Qwen2.5-VL-32B-Instruct-bnb-4bit:

Qwen2.5: model family and version

VL: model purpose (pre-trained for vision-language)

32B: number of parameters (32 billion trainable neural network weights)

Instruct: fine-tuned for instruction following

bnb-4bit: BitsAndBytes 4-bit quantization

Interested in learning more? This article is part of a series on AI fine-tuning. You can read the most recent article on dataset construction here.

5 best product analytics software solutions for product managers

Quadratic — Wed, 26 Mar 2025 22:04:05 +0000

A survey by Forbes revealed that data-driven businesses are 23 times more likely to top their competitors in customer acquisition and about 19 times more likely to stay profitable. This emphasizes the power of data in understanding user behavior, analyzing usage patterns, monitoring feature adoption, and making informed strategic decisions.

Extracting valuable data to enhance user experience and drive revenue can be a lot of work for product managers. This is where product analytics tools become essential. These tools help to make data-driven business decisions and improve overall productivity. By using these tools to automate your workflow, product managers can efficiently analyze user data, identify what works, and focus on strategies that drive engagement and boost growth.

As a product manager, recognizing the need for product analytics software is just the first step; choosing the right tool for your project is equally important. With so many options available, finding the best fit can be challenging.

In this blog post, we’ll discuss the 5 best product analytics software solutions, each designed to help you transform data to business decisions with ease. From real-time dashboard creation to AI-powered insights, these tools offer unique features to enhance your analytics strategy.

1. Quadratic

Quadratic is an AI-powered spreadsheet designed to streamline product management analytics and help project managers transform raw data to actionable insights effortlessly. What sets Quadratic apart as one of the best product analytics tools is its speed in generating insights, thanks to its native AI integration. Product managers can use Quadratic AI to gain valuable insights from user data, identify patterns in user behavior, and develop data-driven growth strategies.

It offers a centralized platform to view, analyze, and visualize user data, eliminating the need to juggle multiple tools to achieve the desired analytics result. This allows product managers to focus on strategies that drive product growth rather than manually cleaning, analyzing, and visualizing data. Its familiar spreadsheet interface makes it easy for non-technical product managers to self-serve user analytics.

Beyond AI assistance, Quadratic offers helpful features such as real-time collaboration with team members, direct database connectivity, built-in support for modern programming languages like Python, SQL, and JavaScript, and advanced data visualizations, all while maintaining enterprise-grade security with SOC 2 and HIPAA compliance. Let’s explore these features in detail:

Features of Quadratic

AI assistance: Whether you’re a technical user or citizen developer, Quadratic’s built-in AI lets you derive insights in seconds by asking questions using simple text prompts. Leveraging LLMs for data analysis reduces the time spent manually analyzing usage patterns and understanding user behavior. It also encourages data-driven decision-making since you gain insights from your data without external factors. Here’s an example:

In this example, I generated a sample user data table containing key data points like join date, last active date, engagement score, feature package, and device type. With these data points, I was able to uncover valuable insights pertaining to user behavior. For instance, I asked Quadratic’s AI (Claude 3.7 Sonnet, in this case), "What feature package is most used on mobile devices?". In response, it instantly generated a summary of the output, provided the corresponding code, and created a separate table analyzing the results. Here’s the result:

The table shows a breakdown of the feature packages used exclusively by mobile users. Quadratic AI generates Python code for accurate analysis, allowing you to derive a wide range of insights by simply asking questions as though you were having a face-to-face interaction with your data. For instance, if I want to compare user engagement between desktop and mobile devices, I only have to ask.

Here, I asked Quadratic AI if users spend more time on the app using their mobile or desktop devices. It compares user distribution across mobile and desktop devices and the average time spent on each device. Learn how you can summarize data tables.

Collaboration: Quadratic offers a collaborative environment where teams can collaborate in real-time and generate insights simultaneously. You can easily invite other team members and control their access to a particular file.

Multiple data sources: Quadratic enables seamless integration with databases, APIs, and raw data. You can effortlessly pull data from multiple sources, view it within Quadratic’s interface, analyze user actions, and generate valuable insights from your data. For instance, you can import data from Mixpanel via API into Quadratic to conduct more advanced analyses beyond Mixpanel’s built-in capabilities.
Data visualization: Visualizations make it easier for product managers to understand their data. With Quadratic AI, you can instantly generate various charts—such as bar charts, scatter plots, line charts, and area charts—based on your data. Let’s see how we can visualize the comparison between mobile users and desktop users based on the total time spent:

Quadratic AI automatically generates charts based on your text prompts. Simply describe how you want to present your data, and Quadratic will write Python code with built-in Plotly libraries to create an interactive chart.

Performance: Built on a modern web app architecture, Quadratic delivers high-performance product usage analytics, providing a smooth and responsive experience even when handling large datasets.
Pricing: Quadratic is free for individuals, with limits to AI prompts you can send. For teams of three or more, a plan is available at $18/month, offering 10-20x higher AI limits, unlimited sharing, and access to a shared team workspace.

2. Mixpanel

Mixpanel is a product analytics tool that provides product managers with real-time insights into user behavior, helping them make more informed decisions with instant access to their data. It enables detailed event tracking, such as button clicks and page views, across both web and mobile applications. It offers a centralized platform that consolidates product analytics metrics such as data views, event tracking, and cross-platform segmentation.

One of Mixpanel’s main drawbacks is its initial setup process, as you have to manually create user events in your application and send them to Mixpanel’s servers for tracking, which could be time-consuming. It has a steep learning curve for beginners, and maximizing certain features may require some expertise.

Mixpanel offers many features, including event tracking, real-time insights, funnel analysis, integration with third-party tools, and session replays. Let’s discuss some of these features:

Features of Mixpanel

Detailed event tracking: Mixpanel enables tracking of user interactions such as button clicks, form submissions, and page views. Product managers can analyze actions that drive the most engagement from users and leverage that accordingly.
Real-time insights: Product managers can access real-time dashboards and reports and identify usage patterns instantly.
Session replays: This feature allows you to watch session replays of individual users, providing more context into user actions and helping to identify what truly drives user events.
Integrations: Mixpanel integrates with platforms such as Quadratic, Google Cloud, Zoho, and Zapier.
Pricing: Mixpanel offers a free plan with a limit of 1 million monthly user events. For usage beyond this, the paid Growth plan charges $0.00028 per event.

3. Amplitude

Amplitude is a leading product analytics tool that offers deep insights into user behavior, helping businesses make tailored decisions for product growth. It is known for its user segmentation and behavioral cohorts, which segment users based on their actions on the app and engagement score. Product managers can better understand their users’ journey with instant visualizations, automated reports, and real-time insights.

Amplitude’s product analytics dashboard allows tracking of key growth metrics such as conversion, retention, and engagement. Given its steep learning curve and wealth of features, Amplitude is often perceived as an overkill for small teams that are only looking to track basic user events. To flatten the learning curve, it offers some built-in customizable analytics templates, allowing product managers to access key metrics instantly. Amplitude users can also integrate their data with Quadratic for additional insights.

Features of Amplitude include behavioral cohorts, retention analysis, funnel analysis, event segmentation, and session replays. Let’s discuss some of these.

Features of Amplitude

Behavioral cohorts: This feature identifies users with similar behavior and automatically groups them into a cohort. This helps to easily connect user behavior to specific business outcomes.
Session replay: Amplitude allows product managers to integrate session replay into their applications, providing a comprehensive view of user behavior. This helps to identify friction and optimize user experience.
Retention analysis: This feature provides product managers with insights into their active users, helping them understand how frequently they return and the reasons behind their engagement. It also offers strategies for identifying the most relevant retention metrics tailored to a specific business.
Funnel analysis: Amplitude’s robust funnel analysis feature provides insights into your product’s onboarding funnel, helping product managers identify drop-off points and how they can optimize conversion rates.
Pricing: Amplitude offers a free plan with limited features, while its paid plan for small teams starts at $49 per month(billed annually).

4. Heap

Heap is a product analytics software renowned for its automatic data capture, which records all user interactions on a website or app without requiring manual setup. Unlike Amplitude and Mixpanel, which need technical configuration for event tracking, Heap automatically tracks user actions such as button clicks, form submissions, and page views.

This automation provides quick access to analytics for product managers without additional setup. Heap also offers advanced data science capabilities, which enable teams to uncover friction points in user behavior. Similar to Quadratic, Heap is relatively easy to use, thanks to its automatic data capture feature. It integrates with platforms such as Salesforce, HubSpot and Shopify.

Its key features include session replay, funnel analysis, heatmaps, user cohorts, and customizable charts. However, its chart customization options are somewhat limited compared to other tools.

Features of Heap

Automatic data capture: Heap’s automatic data capture feature eliminates the need for manually tagging user events, which saves time and improves productivity.
Heap Illuminate: Heap offers a data science layer that allows product managers to analyze users’ behavioral data and automatically generate all sorts of insights, including actions they haven’t been following. This feature is called Heap Illuminate.
Heatmaps: Heatmaps enable product managers to get visual insights into what users engage with and what they ignore, with no manual event creation or tagging required.
Cohorts: Like Amplitude, Heap also allows product managers to create user cohorts based on the actions they take on their website or app.
Pricing: Heap offers a free plan that offers up to 10,000 monthly sessions. For additional features(such as chart customization and email support), it provides paid plans, but pricing varies based on the number of user sessions. To get an exact estimate, users must sign up for a quote.

5. Google Analytics

Google Analytics is a popular player in the product analytics game as it remains the de facto application for analyzing user activity on websites and apps. It provides valuable metrics such as page views, conversion rates, user locations, and per-click data. With Google Analytics, product managers can gain insights into where their users come from and the devices they use.

While Google Analytics performs excellently in tracking website traffic and funnel metrics, it falls short in the aspect of user behavior analysis. Unlike other product analytics tools, it does not offer features like session replays, user journeys, or heatmaps. As a result, product managers cannot fully understand how users interact with their products unless they integrate Google Analytics with additional session replay tools.

Product managers seeking a comprehensive insight into what is happening on their website will find Google Analytics highly valuable. However, for deeper insights into why and how users behave, other analytics tools with advanced in-app event tracking may be more suitable. Key features of Google Analytics include traffic analysis, real-time analytics, predictive capabilities, funnel analysis, and seamless integrations with various platforms.

Features of Google Analytics

Real-time analytics: Changes to user data happen in real-time. Google Analytics allows you to monitor activity on your website or app as it happens.
Predictive capabilities: Google’s machine learning models provide product managers with predictive insights into user behavior, allowing them to anticipate future actions and tailor strategies for specific audiences. This sets the foundation to optimize conversion rates and enhance customer retention.
Traffic analysis: This feature helps you easily identify where your visitors are coming from, whether through search, direct traffic, paid ads, or social media. It also provides insights into which channels need better marketing efforts to drive more engagement.
Integrations: Google Analytics is known for its seamless integration with various platforms, including those within the Google ecosystem (Google Ads, Google Optimize, Google BigQuery), BI tools (Power BI, Tableau, Looker), and CRMs (HubSpot, Salesforce).
Pricing: Google Analytics offers a free version that includes core features, which makes it the most cost-effective solution for most businesses. It also has a paid version of $50,000 annually that comes with more advanced features.

Summary: Choosing the best product analytics software

Data analytics for product managers is essential for driving growth and making data-driven decisions. In this blog post, we explored five analytics tools for product managers. Each tool has its strengths, and choosing the best product analytics software depends on your specific needs. Here’s a summary to help you make the right choice:

For detailed in-app event tracking and deeper insights into user behavior, choose Mixpanel or Amplitude.
For a friendlier platform with automated data capture and effortless event tracking, Heap is the best option.
For a cost-effective solution with real-time analytics and comprehensive traffic analysis, Google Analytics stands out.
For AI-powered product data analytics, a familiar spreadsheet interface, real-time collaboration, direct connection to multiple data sources, and advanced data visualizations on your user data, Quadratic is the ideal choice.

Ready to experience the future of analytics and make data-driven decisions for your business or product? Try Quadratic for free today.

Quadratic: The AI-Powered Spreadsheet for Modern Teams

Quadratic — Tue, 18 Mar 2025 22:56:15 +0000

Spreadsheets have been around for decades, but they’ve largely remained the same: rows, columns, and a handful of formulas. We built Quadratic to change that, making it much easier to get insights from data. It’s a modern spreadsheet that integrates AI natively, connects directly to your data, and supports not just formulas, but also SQL, Python, and JavaScript—all in one collaborative environment.

Why Quadratic?

We believe spreadsheets should be as powerful as any modern data tool while remaining accessible to everyone. Typical desktop-based solutions can get unwieldy, especially when you need real-time collaboration or advanced analytics. Quadratic eliminates those barriers with:

No setup or downloads: Open Quadratic in your web browser and start working instantly with anyone in real-time.
Democratized data access: Empower your team to make data-driven decisions by connecting directly to your database or any public API.
All-in-one: For quick insights, you don’t need five different tools. Quadratic is a database connector, AI-powered IDE, and BI tool with a familiar spreadsheet interface.

What makes Quadratic different?

AI-powered - Quadratic supports the latest LLMs like Claude 3.7 Thinking, Claude 3.5 Sonnet, and GPT-4o, which excel at writing and refining code. Let the AI generate formulas, SQL queries, or even Python/JavaScript for you. Then, review and adapt the code yourself for full transparency.
Support for SQL, Python & JS - Go beyond traditional spreadsheet formulas. If you need complex logic or want to query large datasets, Quadratic’s integrated code cells let you write SQL, Python, or JavaScript directly in your spreadsheet. It’s perfect for building robust models or connecting to external APIs in just a few lines of script.
Live connections & API integrations - Easily link Quadratic to your databases (e.g., PostgreSQL, MySQL) or third-party data sources. With direct, live connections, you can pull in real-time updates or run SQL queries without ever leaving your spreadsheet.
Modern performance stack - Quadratic is built on Rust, WASM, and WebGL, which means it’s snappy, secure, and scalable. Sometimes, the little interactions throughout your workflow make your life easier.

Ready to see it in action?

If you’ve ever been frustrated by clunky spreadsheets that can’t handle real-time collaboration or advanced analytics, give Quadratic a spin. There’s nothing to install: just open your browser, connect your data, and let our AI boost your workflow.

Try Quadratic for free and discover a smarter, faster way to work with data!

Making API requests from your spreadsheets

Quadratic — Fri, 14 Mar 2025 20:08:46 +0000

With Quadratic, you can connect APIs to spreadsheets in a programmatic way to access live data from any data source exposed via an API endpoint. You can read from APIs with GET requests and write to APIs with POST requests, all directly from Quadratic's native Python experience.

Connecting your spreadsheet to your APIs is easy. We'll start with the simplest possible API, with no credentials or parameters, and then move to one with credentials and parameters.

Step 1: identify your API endpoint

In this example we're using a free sample API from JSONPlaceholder.

The endpoint is:

https://jsonplaceholder.typicode.com/users

We can view what some of that sample data looks like in our browser.

Step 2: build query and view in spreadsheet

This is the simplest possible query as there are no parameters, only a URL that returns our data.

import requests
import pandas as pd

# make request
x = requests.get('https://jsonplaceholder.typicode.com/users')

# go from JSON response to DataFrame
df = pd.DataFrame.from_dict(x.json())

# display DataFrame in the sheet
df

This returns our results, which are tabular in the spreadsheet.

Step 3: manipulate the DataFrame

From here, we can manipulate the DataFrame, doing statistics, visualizations, or selecting individual pieces of data. Here are just a few examples:

Select a column

df['username']

This returns only the username column to the sheet.

Summarize the column's data

We can do analytics on our selected data. In this example, we count the number of times a username starts with a specific letter. This would have taken a bit to figure out, so instead, we can just ask the AI assistant in the console. On the first try, the AI performed an answer that worked great.

Make a chart

With our summarized data from the last step, we can make a simple chart showing each letter's occurrence rate, such as the starting letter in a username.

# import plotly
import plotly.express as px

# create your chart type, for more chart types: https://plotly.com/python/
fig = px.bar(df, x = letter, y = frequency)

# make chart prettier
fig.update_layout(
    plot_bgcolor="White",
    height=700
)

# display chart
fig.show()

API requests with parameters (plus POST instead of GET)

For this next example, we connect to Exa AI, a search AI API. The same steps are used above, just with query parameters, changing our API request slightly. It's also a POST instead of a GET request, meaning we'll POST data, and then they'll respond with an answer to what we're looking for. POST requests are also a common way to write data to external sources, where the response is a confirmation of whether or not the data was successfully received. In this case, we're sending data and receiving an answer using our POST request.

In this example, we build our request from parameters set in the sheet so we can dynamically make requests to the API based on sheet data.

Building API request with parameters

We need to follow the parameters specifications in the API's documentation. In the Exa example, we must provide the query, category, and number of results as the payload; as headers, we need to minimally supply our API key and the data type we will receive. See the example below for how this works.

import requests
import pandas as pd

# API URL
url = "https://api.exa.ai/search"

# payload
payload = {
    # adjusts our query to an optimized prompt
    "useAutoprompt": True,
    # returns 10 results
    "numResults": 10,
    # category of response we want
    "category": "company",
    # question we want to ask, read from spreadsheet cell (0,3)
    "query": cell(0,3)
}

# headers
headers = {
    # what data format we'll accept
    "accept": "application/json",
    # what data format we're sending
    "content-type": "application/json",
    # API key
    "x-api-key": "your_api_key_here",
}

# send request
response = requests.post(url, json=payload, headers=headers)
# turn response into json format we can shove into DataFrame
response.json()['results']
# put request into DataFrame
df = pd.DataFrame.from_dict(response.json()['results'])
# dislay DataFrame to sheet
df

Summary

API requests are a powerful way to connect to live data sources from your spreadsheet—query all sorts of data sources directly from Python in Quadratic to level up your analytics.

Feel free to contact us and let us know how or what we can improve in Quadratic. User feedback is what directly guides our product roadmap.

Building a community-led product roadmap

Quadratic — Thu, 13 Mar 2025 21:18:27 +0000

One of the best things about product planning at Quadratic is our extremely opinionated user base. We send our active users personal emails asking what they'd improve in Quadratic, and their responses are massive, thoughtful, multi-paragraph emails from which we can glean new ideas and patterns. We also receive feedback through feedback forms in the application, feedback from prospects in feedback and sales calls, and developer input on GitHub as a source-available project.

Simultaneously, as the builders of Quadratic, we are opinionated on where the best spreadsheets and data tools are needed to take their products. The product team is tasked with digesting opinions and feedback we're hearing the most, which is the most valuable, what beliefs do we have indpendent of feedback, and when is the scope of a feature too large to justify building just because a user says it's important. Below is a look into our core beliefs and the user feedback that has directly influenced our roadmap.

Note: our roadmap is a living, public project on GitHub. Issues are added and removed daily to improve Quadratic.

What are the core beliefs represented in our product roadmap?

When our founding team endeavored to build a better spreadsheet, some core beliefs informed the initial product vision. These beliefs were informed by our team's experiences doing analytics in past companies and interviewing data people in and out of the network. These beliefs led to the first versions of Quadratic and our pre-seed round in 2022.

Some of these beliefs include:

Most organizations pursue automation in or out of their spreadsheets, often powered by Python.
Citizen developers are spreading throughout organizations, disrupting all sorts of prior workflows; they should be empowered by better tooling.
Despite the rise of code and data tooling, most analytics end up in a spreadsheet; export to CSV is the most commonly used feature in many BI tools.
Developers are no longer sitting back and using the tools assigned to them; they're testing, building, and self-hosting modern tools that then spread like wildfire across their organizations.
Developers have pull in orgs, but when they leave, the tools they leave behind are hard to maintain.
Existing spreadsheets have performance issues (data limitations, clunky UX, outdated feature sets, etc.) Modern technology enables us to build a better sheet.
Existing spreadsheets aren't a great collaborative experience. In the age of great collaborative tools like Figma and Notion, this should change.
Quadratic is growing in the age of AI. Analytics will be done in a spreadsheet, empowered by AI, not in a chat interface.
Open-source and source-available projects enhance users' trust and help build community.

These beliefs informed a clear initial roadmap, highlighted by the following features, most of which exist in Quadratic today:

A performant, web-based spreadsheet that runs locally and smoothly at 60fps with Rust + WASM + WebGL.
The ability to handle millions of rows of data directly in the browser.
Python built into the spreadsheet in an intuitive way that feels native to the spreadsheet, with excellent developer experience.
Charts and graphs (currently via Python + Plotly, expanding support in the future, native eventually).
Real-time collaboration with your team in the spreadsheet.
AI built-in to help users write code and build better analytical functions.
Live connections to wherever a user's data lives via SQL, Python requests, or custom data connectors.
AI that enables automating significant percentages of a typical user workflow.
Self-hosted and on-prem solutions for enterprise organizations.
(Long-term WIP) the best core spreadsheet experience available.

How does the Quadratic community inform this roadmap?

GitHub

We receive feedback in various formats. One of our core sources is GitHub. As the only modern spreadsheet building with our source code available, we're privileged to have developers of all kinds contribute both feedback and code. The feedback we get through GitHub is developer-focused, usually centered around technical questions, deployment options, open-source/contributing specifics, etc.

In-app feedback forms

We present contact forms in various places for users to tell us what they love and what they don't. This exists both on our marketing website and in-app. Forms usually receive the direct papercuts users experience in-app. These are usually bugs that we push to the top of our roadmap and resolve within a few hours, sometimes within a few minutes.

We've implemented many QA tests, monitoring, and other key infrastructure to minimize bugs, but it's not uncommon for something to slip through the cracks. Users will always push your product to the limit and find the places you didn't think to go in testing. Once they do, we implement the fix and add their edge case to the test suite. This class of user feedback is an essential element for improving Quadratic.

Prospects

During sales calls, there are sometimes obvious things that many enterprise users can't live without. Some of those features they've told us about include:

SQL data connectors (see all connections).
SOC 2 compliance for enterprises (we're SOC 2 and HIPAA compliant now).

Solicited feedback

We spend a lot of time with personalized outreach to users, asking what they think can be improved. It's a lot of effort and time spent digesting feedback, but it's clearly worth it to us. Where some products might get crickets any time they try to solicit feedback, our community is extremely passionate about letting us know where we need to improve. It's especially obvious what to build when we hear the same feedback repeatedly.

In these situations, we have no choice but to bump that feedback to the top of our roadmap. Some features we shipped recently that were constantly getting asked for include:

Inline formulas instead of a multi-line editor by default.
Relative references.

Some features we are actively working on were also consistently asked for by our general community.

DateTime support.
Cell ranges in Python.

Summary

Roadmap construction at Quadratic is all about balancing our obvious user feedback with our opinionated view on what will make Quadratic the best possible spreadsheet.

If you want to influence our roadmap, feel free to contact us and let us know how or what we can improve in Quadratic. As you've learned above, user feedback is the primary driver for what we build next.

Using JavaScript in a spreadsheet

Quadratic — Mon, 10 Mar 2025 21:51:20 +0000

Our goal with Quadratic is to support many major programming languages. We started with Python as the most popular language for data science. Along the way, we've continued improving that experience to build the best (and first) Python experience in a spreadsheet.

We're excited to now lead the way with JavaScript! As the world's most popular programming language, we expect plenty of interest in combining JavaScript with the world's most popular tool for interfacing with data - the spreadsheet.

This article discusses how to make your spreadsheets even more powerful by leveraging JavaScript in Quadratic.

Step 1: create a JavaScript code cell

Press / to code.

Once your editor is open, there are a few ways to get started. Highlighted in the image below are:

AI assistant: prompted to only help with JavaScript questions while in the JS editor
Code snippets: quick snippets to help you get started
Console: for your code outputs, errors, etc.
Play/pause/etc: for running code and pausing/stopping long-running executions

Step 2: reference data

Reference data from the spreadsheet to get data into JavaScript that currently lives in the spreadsheet.

The following is the Quadratic way to reference data in the sheet from JavaScript.

// Reference a single value from the sheet; replace x,y with coordinates
let data = await cell(x, y);

// Or reference a range of cells (returns an array), replace x's and y's with coordinates
let data = await cells(x1, y1, x2, y2)

// Reference cell or range of cells in another sheet
let data = await cells(x1, y1, x2, y2, sheet_name)

Step 3: return data to the sheet

You've already referenced data from the sheet into JavaScript, now learn how to return data back to the sheet from JavaScript.

In our JavaScript implementation, your return statement is what gets returned to the sheet. This is unlike our Python implementation where the last line is always returned by default.

You can return single values, 1-d arrays, multi-dimensional arrays, charts, and more.

// displays vertically - 1x4
let x = [1,2,3,4]

// displays horizontally - 4x1
let y = [[1,2,3,4]]

// 4x2
let z = [[1,2,3,4],[1,2,3,4]]

// return statement is what gets printed to the sheet
return x;

Step 4: use libraries

Through ESM modules you can use the JavaScript libraries you're familiar with. Below is an example of using charts.js to build a chart in Quadratic using JavaScript.

Sample import below:

import Chart from 'https://esm.run/chart.js/auto';

Step 5: use Fetch to make an API request

Make an API request in JavaScript by using the JavaScript method fetch(). Below is an example querying a sample API.

// query using fetch
let res = await fetch("https://jsonplaceholder.typicode.com/todos/1");
// put results in json
let json = await res.json();

// print result to console
console.log(json);

// format result json into array
let as = [Object.keys(json), Object.values(json)]

// return array to sheet
return as;

We're excited to release JavaScript as an experimental feature, adding an additional programming language to your toolbelt. As this is an experimental feature we expect there to be some rough edges - feel free to contact us and let us know how or what we can improve in JavaScript or Quadratic as a whole. User feedback is what directly guides our product roadmap.