DEV Community: Mohsin Raza

Real Examples of Business Intelligence in Action

Mohsin Raza — Fri, 04 Jun 2021 04:18:35 +0000

Business intelligence (BI) can add value to almost any business process, creating a comprehensive view and empowering teams to analyze their own data to find efficiencies and make better day-to-day decisions.

Digital transformation is now seen as a key strategic initiative and business intelligence tools have evolved to help companies make the most of their data investments. The response is the rise of modern business intelligence platforms that support data access, interactivity, analysis, discovery, sharing, and governance. While there are some great books about business intelligence that detail practical applications, this article will show off how some specific, big-name companies leverage modern business intelligence platforms.
Here are 5 real-world examples of business intelligence platforms in action.

1. HelloFresh centralized digital marketing reporting to increase conversions

Photo by Marvin Meyer on Unsplash
Company: HelloFresh
Problem: Digital marketing reporting was time-intensive, manual, and inefficient.
Solution: For meal kit company HelloFresh, a centralized business intelligence solution saved the marketing analytics team 10–20 working hours per day by automating reporting processes. It also empowered the larger marketing team to craft regional, individualized digital marketing campaigns.
Based on aggregate analyses of customer behavior, HelloFresh created three buyer personas to guide their efforts. Being able to see and track real-time data means the team can react to customer behaviors and optimize marketing campaigns. As a result, they saw increased conversion rates and improved customer retention.

2. REI increased membership rates for co-op retailer

Company: REI
Problem: Difficulty tracking membership metrics with 90 terabytes of data.
Solution: In this example, Outdoor retail co-op REI uses a business intelligence platform to analyze their co-op membership. Co-op members contribute to REI’s account for more than 90 percent of purchases with the retailer, so it is critical to track metrics like acquisition, retention, and reactivation. All of this information equates to over 90 terabytes of data. The ability to parse all of this data means that operations teams can determine whether to invest more in brick-and-mortar retail or digital experiences for their members.

This leads to greater customer satisfaction and positive associations with the brand.
“We’ve seen a complete turnaround in 2017 with new member acquisition,” observed Clinton Fowler, Director of Customer and Advanced Analytics at REI.
The team also uses their BI platform to analyze customer segmentation, which helps inform decisions like shipping methods, member lifecycle management, and product category assortments.
More reading: Read about the top 5 retail analytics trends.

3. Coca-Cola Bottling Company maximized operational efficiency

Company: Coca-Cola Bottling Company (CCBC), Coca Cola’s largest independent bottling partner
Problem: Manual reporting processes restricted access to real-time sales and operations data.
Solution: Coca-Cola’s business intelligence team handles reporting for all sales and delivery operations at the company. With their BI platform, the team automated manual reporting processes, saving over 260 hours a year — more than six 40-hour work weeks.

Report automation and other enterprise system integrations put customer relationship management (CRM) data back into the hands of sales teams in the field through mobile dashboards that provide timely, actionable information and a distinct competitive advantage.
A self-service BI implementation fosters more effective collaborations between IT and business users that maximize the expertise of participants. Analysts and IT can focus on big-picture strategy and long-term innovations such as enterprise data governance rather than manual research and reporting tasks.
More reading: Tips to Succeed with Big Data in Data Science.

4. Chipotle created a unified view of restaurant operations

Company: Chipotle
Problem: Disparate data sources hindered teams from seeing a unified view of restaurants.
Solution: Chipotle Mexican Grill is an American restaurant chain with more than 2,400 locations worldwide. Chipotle retired their traditional BI solution for a modern, self-service BI platform. This allowed them to create a centralized view of operations so they can track restaurant operational effectiveness at a national scale.

Now that staff has more access to data, the speed of report delivery for strategic projects has tripled from quarterly to monthly and saved thousands of hours. “This was the ticket to take all metrics and understanding to that next level,” explained Zach Sippl, Director of Business Intelligence.
More reading: Why you need a BI platform and how to choose one.

5. Des Moines Public Schools identifies and helps at-risk students

Organization: Des Moines Public Schools
Problem: Manual Excel reporting meant administrators couldn’t see up-to-date data like attendance, preventing timely intervention.
Solution: Des Moines Public Schools (DMPS) used advanced analytics to improve dropout intervention rates and better understand the impact of various teaching methods on individual student outcomes.

The DMPS Research and Data Management team used a multiple linear regression model — nicknamed the dropout coefficient — to weigh student indicators to predict which students might be at risk of dropping out of school. They used a business intelligence platform to leverage the model. Data visualization made it easy for staff to identify individual, at-risk students and get those students the attention they need.
Dashboards set up by the Research and Data Management Team delivered real-time analytics to 7,000 DMPS teachers and staff so they could adapt and intervene sooner, dramatically improving the intervention success rates. The real-time analytics were supported by five years of historical data. This meant that staff could dig into historical data on the spot to validate insights on current students.
https://medium.com/nerd-for-tech/5-real-examples-of-business-intelligence-in-action-866b33503bc0

Q&A for Time Series Analysis/Forecasting

Mohsin Raza — Thu, 13 May 2021 18:15:20 +0000

1. What is Time series analysis?

Time Series is a series of observations taken at specified time intervals usually equal intervals. Analysis of the series helps us to predict future values based on previously observed values. In the Time series, we have only 2 major variables, Time & the variable we want to forecast.

2. Why & where Time Series is used?

Time series data can be analyzed in order to extract meaningful statistics and other characteristics. It’s used in at least 4 scenarios:

a) Business Forecasting
b) Understand past behavior
c) Plan the future
d) Evaluate current accomplishment

3. When shouldn’t we use Time Series Analysis?

We don’t need to apply the Time series in at least the following two cases:
a) The dependant variable(y) (that is supposed to vary with time) is constant. Eq: y=f(x)=4, a line parallel to x-axis(time) will always remain the same.
b) The dependant variable(y) represents values that can be denoted as a mathematical function. Eq: sin(x), log(x), Polynomials etc. Thus, we can directly get value at some time using the function itself. No need for forecasting.

4. What are the components of the Time Series?

There are fourth components:
a) Trend Upward & downward movement of the data with time over a large. period of time. Eq: Appreciation of Dollar vs rupee.
b) Seasonality Seasonal variances. Eq: Ice cream sales increases in Summer only.
c) Noise or Irregularity Spikes & troughs at random intervals.
d) Cyclicity Behavior that repeats itself after a large interval of time, like days months & years, etc.

5. What is Stationarity?

Before applying any statistical model on a Time Series, the series has to be stationary, which means that over different time periods,
a) It should have a constant mean.
b) It should have constant variance or standard deviation.
c) Auto-covariance should not depend on the time.

6. Why does Time Series(TS) need to be stationary?

It is because of the following reasons:
a) If a TS has a particular behavior over a time interval, then there’s a high probability that over a different interval, it will have the same behavior, provided TS is stationary. This helps in forecasting accurately.
b) Theories & Mathematical formulas are more mature & easier to apply for as TS which is stationary.

7. Tests to check if a series is stationary or not.

There are 2 ways to check for Stationarity of a TS:
a) Rolling Statistics →Plot the moving avg or moving standard deviation to see if it varies with time. It's a visual technique.
b) ADCF Test →Augmented Dickey-Fuller test is used to gives us various values that can help in identifying stationarity. The Null hypothesis says that a TS is non-stationary. It comprises Test Statistics & some critical values for some confidence levels.
If the Test statistics are less than the critical values, we can reject the null hypothesis & say that the series is stationary. THE ADCF test also gives us a p-value. Acc to the null hypothesis, lower values of p is better.

8. What is the ARIMA model?

ARIMA(Auto-Regressive Integrated Moving Average) is a combination of 2 models AR(Auto-Regressive) & MA(Moving Average).
It has 3 hyperparameters — P(auto-regressive lags),d(order of differentiation), Q(moving avg.) which respectively come from the AR, I & MA components. The AR part is the correlation between prev & current time periods. To smooth out the noise, the MA part is used. The I part binds together the AR & MA parts.

9. How to find the value of P & Q for ARIMA?

We need to take the help of ACF(Auto Correlation Function) & PACF(Partial Auto Correlation Function) plots. ACF & PACF graphs are used to find the value of P & Q for ARIMA. We need to check, for which value in the x-axis, graph line drops to 0 in the y-axis for 1st time.
From PACF(at y=0), get P
From ACF(at y=0), get Q

10. What Is the ADCF test?

In statistics and econometrics, an augmented Dickey-Fuller test (ADF) tests the null hypothesis that a unit root is present in a time series sample. The alternative hypothesis is different depending on which version of the test is used but is usually stationarity or trend-stationarity. It is an augmented version of the Dickey-Fuller test for a larger and more complicated set of time series models.
The augmented Dickey-Fuller (ADF) statistic, used in the test, is a negative number. The more negative it is, the stronger the rejection of the hypothesis that there is a unit root at some level of confidence.
p-value(0<=p<=1) should be as low as possible. Critical values at different confidence intervals should be close to the Test statistics value.

11. What is Exponential Smoothing?

Exponential smoothing is a rule of thumb technique for smoothing time series data using the exponential window function. Whereas in the simple moving average the past observations are weighted equally, exponential functions are used to assign exponentially decreasing weights over time. It is an easily learned and easily applied procedure for making some determination based on prior assumptions by the user, such as seasonality. Exponential smoothing is often used for the analysis of time-series data.
The raw data sequence is often represented by xt beginning at time t=0, and the output of the exponential smoothing algorithm is commonly written as st, which may be regarded as the best estimate of what the next value of xx will be. When the sequence of observations begins at time t=0, the simplest form of exponential smoothing is given by the formulas:
s0 = x0
st = α∗xt+(1−α)∗st−1 , t>0
where α is the smoothing factor, and 0<α<1.

12.What is Exponential decay?

A quantity is subject to exponential decay if it decreases at a rate proportional to its current value. Symbolically, this process can be expressed by the following differential equation, where N is the quantity and λ (lambda) is a positive rate called the exponential decay constant:
dN/dt = −λN
The solution to this equation (see derivation below) is:
N(t) = N0∗e−λt
where N(t) is the quantity at time t, and N0 = N(0) is the initial quantity, i.e. the quantity at time t = 0.
Half-Life: is the time required for the decaying quantity to fall to one-half of its initial value. It is denoted by t1/2. The half-life can be written in terms of the decay constant as:
t1/2=ln(2)/λ

Thanks for reading my article please share your thoughts.
Follow me on LinkedIn