🧠 “Data Formats: The Avengers of Analytics” (And Who’s the Real Hero?)

Prabakaran S R — Mon, 06 Oct 2025 14:18:01 +0000

So you’ve got a dataset — a few names, marks, maybe your friend’s secret crush score 😏 — and you’re told to store it. But how?

That’s where our six heroes of data storage enter:

CSV, SQL, JSON, Parquet, XML, and Avro.

Each has its own personality — some simple, some complicated, and some that just exist to confuse you at 3 AM.

Let’s meet the squad 👇

🧩 Our Mini Dataset

Let’s keep it simple:

Name	Register Number	Subject	Marks
Praba	21CS001	Cloud Data	92
Alex	21CS002	Big Data	88
Sam	21CS003	AI	95

1️⃣ CSV — The Simplicity King 👑

CSV (Comma-Separated Values) is like the humble text file that stores your data line by line, separated by commas.

No fancy metadata, no drama. Just pure simplicity.

csv

Name,Register Number,Subject,Marks Praba,21CS001,Cloud Data,92 Alex,21CS002,Big Data,88 Sam,21CS003,AI,95

🗣️ “I’m small, fast, and open in Excel. What else do you need?” — CSV, probably

2️⃣ SQL — The Organized Perfectionist 🧮

SQL stores data in structured tables with rows and columns, like a well-maintained hostel attendance sheet.
It loves order and rules — “no duplicate primary keys, please.”

`CREATE TABLE Students (
Name VARCHAR(50),
RegisterNumber VARCHAR(10),
Subject VARCHAR(50),
Marks INT
);

INSERT INTO Students VALUES
('Praba', '21CS001', 'Cloud Data', 92),
('Alex', '21CS002', 'Big Data', 88),
('Sam', '21CS003', 'AI', 95);
`
🗣️ “I believe in relationships… relational databases, to be precise.” — SQL

3️⃣ JSON — The Developer’s BFF 💻

JSON (JavaScript Object Notation) is loved by APIs and front-end devs everywhere.
It’s structured yet flexible — perfect for sending data between systems.
[ { "Name": "Praba", "RegisterNumber": "21CS001", "Subject": "Cloud Data", "Marks": 92 }, { "Name": "Alex", "RegisterNumber": "21CS002", "Subject": "Big Data", "Marks": 88 }, { "Name": "Sam", "RegisterNumber": "21CS003", "Subject": "AI", "Marks": 95 } ]
🗣️ “I speak fluently with JavaScript. JSON.stringify me!”

4️⃣ Parquet — The Speed Demon 🏎️

Parquet is a columnar storage format used in Big Data tools like Spark and Hadoop.
It’s highly compressed and optimized for reading specific columns fast — like “give me all the marks” instead of scanning the whole file.

Example?
Parquet isn’t human-readable (that’s the point!), but if it were:

Columns: Name → ["Praba", "Alex", "Sam"] RegisterNumber → ["21CS001", "21CS002", "21CS003"] Subject → ["Cloud Data", "Big Data", "AI"] Marks → [92, 88, 95]
🗣️ “I’m not pretty, but I’m fast. Ask any data engineer.”

5️⃣ XML — The Drama Queen 📜
XML (Extensible Markup Language) loves tags — open tags, close tags, nested tags… it’s basically HTML’s over-serious cousin.
<Students> <Student> <Name>Praba</Name> <RegisterNumber>21CS001</RegisterNumber> <Subject>Cloud Data</Subject> <Marks>92</Marks> </Student> <Student> <Name>Alex</Name> <RegisterNumber>21CS002</RegisterNumber> <Subject>Big Data</Subject> <Marks>88</Marks> </Student> <Student> <Name>Sam</Name> <RegisterNumber>21CS003</RegisterNumber> <Subject>AI</Subject> <Marks>95</Marks> </Student> </Students>
🗣️ “I may be verbose, but at least I have structure!” — XML

6️⃣ Avro — The Efficient Coder 🧠

Avro is a row-based binary format designed for fast serialization and compact storage.
It’s schema-driven and used in Kafka or streaming pipelines.

Human-readable version (simplified):
{ "type": "record", "name": "Student", "fields": [ {"name": "Name", "type": "string"}, {"name": "RegisterNumber", "type": "string"}, {"name": "Subject", "type": "string"}, {"name": "Marks", "type": "int"} ], "data": [ {"Name": "Praba", "RegisterNumber": "21CS001", "Subject": "Cloud Data", "Marks": 92}, {"Name": "Alex", "RegisterNumber": "21CS002", "Subject": "Big Data", "Marks": 88}, {"Name": "Sam", "RegisterNumber": "21CS003", "Subject": "AI", "Marks": 95} ] }
🗣️ “I’m compact, I’m fast, and I don’t need XML’s drama.” — Avro

💬 Written by Prabakaran SR — a caffeine-fueled data explorer trying not to break the pipeline again ☕💾

Hands-On MongoDB with a Games Dataset

Prabakaran S R — Sun, 24 Aug 2025 15:05:25 +0000

In this blog, I’ll share my practical experience using MongoDB with a games dataset. The exercise focused on storing, querying, and analyzing data, helping me understand how NoSQL databases work in real-world scenarios.

Step 1: Setting Up MongoDB

I used MongoDB Compass locally and created a database named gameDB. Compass provides an intuitive interface to interact with the database, visualize collections, and run queries easily.

Step 2: Importing the Games Dataset

I imported a sample games dataset in JSON format containing the following details:

Game ID

Title

Genre

Rating

Review

After importing, I could see all the games and their reviews neatly organized in the collection.

Step 3: Inserting Records Manually

To practice, I inserted 10 new game records manually. This helped me understand how new data can be added into an existing collection in MongoDB.

Step 4: Performing Queries
Top 5 Games by Rating

I identified the top 5 highest-rated games in the dataset. This helped me practice sorting and filtering data to get meaningful insights.

Counting Reviews with “Good”

I analyzed how many reviews contained the word “good.” This demonstrated MongoDB’s ability to search text across documents efficiently.

Fetching Reviews for a Specific Game

I retrieved all reviews for a specific game using its ID. This showcased how to filter data for targeted analysis.

Updating and Deleting Records

I updated the review of a game to reflect new feedback and also deleted a record that was no longer relevant. This demonstrated MongoDB’s data modification capabilities.

Step 5: Exporting Query Results

After performing the queries, I exported the results in JSON and CSV formats. This allows further analysis or sharing with others.

Conclusion

Working with MongoDB and a games dataset was a highly educational experience. Key takeaways include:

Understanding CRUD operations in NoSQL databases

Practicing data querying, sorting, and filtering

Learning to export and share data efficiently

MongoDB proves to be a powerful and flexible database, especially for datasets with dynamic or semi-structured data like game reviews.

DEV Community: Prabakaran S R

🧠 “Data Formats: The Avengers of Analytics” (And Who’s the Real Hero?)

Let’s meet the squad 👇

🧩 Our Mini Dataset

1️⃣ CSV — The Simplicity King 👑

Hands-On MongoDB with a Games Dataset