BOOKSTORE
A local bookstore engaged our team to conduct a comprehensive data analytics assessment of its customer base to uncover key spending drivers and factors influencing eBook subscription adoption. The provided dataset contained detailed customer demographics, transaction history over a defined period, and subscription statuses for the bookstore’s eBook service.
​
The primary objectives of this analysis were:
​​
-
Identifying Customer Spending Drivers – Analyzing demographic and behavioral trends to determine which factors most significantly influence customer spending patterns.
​
-
Understanding eBook Subscription Adoption – Examining conditions that contribute to customer enrollment in the bookstore’s eBook service.
​
-
Predictive Modeling – Implementing machine learning algorithms, including regression and classification models, to accurately predict customer spending behavior and likelihood of eBook subscription adoption.
​
To achieve these objectives, we conducted Exploratory Data Analysis (EDA) to identify correlations and patterns, leveraging SQL and Python for data preprocessing and feature engineering. We curated optimized datasets to train and test machine learning models, ensuring robust predictive performance. This analysis provided actionable insights to help the bookstore refine its marketing strategies, enhance customer engagement, and drive revenue growth.
Initial Dataset EDA
In the above code, we were able to use DataPrepEDA to see initial data point distributions and take a look at the data structure do some basic engineering.
​
DataPrepEDA Report
​
Once we have completed our EDA, we can take the first section of data and apply a linear regression algorithm to predict customer spend and driver further spend analysis.
Based on the results of the linear regression, we were able to determine that Light Gradient Boosting Machine was the best algorithm suited for our purposes, so we deployed that algorithm to create our model. After that, we were able to take the other set of data and focus on the classification algorithm.
Once we were able to deploy our classification model to determine the likelihood of Ebook Subscription adoption, our efforts were focused on aggregating the data with SQL for final analysis, data visualization, and to draw actionable insights for the relevant stakeholders.
Once the data is aggregated into "Final Data", we will load the data into excel and create a dashboard to visualize, analyze, and present our findings to any relevant stakeholders.
This is the layout of the excel dashboard highlighting demographic trends and projects revenue and subscription adoption for the next period. This was built by building columns , dashboards, and pivot tables in excel.