Business Analyst (Student Project): Bookstore Sales Optimization (market basket analysis)
- Paulina y
- Jun 25, 2025
- 1 min read
Applied CRISP-DM and WEKA for market basket analysis
Situation:
A bookstore wanted to boost sales and improve customer satisfaction through better recommendations and store layout. The business needed insights into which book categories are often purchased together, to enable product bundling and more effective cross-selling.
Task:
As a business analyst student, my objective was to discover associations between book purchases using data mining techniques. The main question was: Which combinations of books should be promoted together to increase sales?
Action:
I used the CRISP-DM framework and prepared the dataset by converting numeric attributes to nominal to enable association rule mining. The Apriori algorithm was applied in WEKA, with parameters adjusted (confidence reduced from 0.9 to 0.75 and rule limit set to 4) to generate more meaningful association rules. I removed irrelevant attributes like ID to focus on purchase patterns.
Result:
Identified strong associations between youth books, cookbooks, and children’s books with a high confidence level of 0.86.
Delivered actionable insights that allowed for optimized bundling, recommendation systems, and in-store layout strategies.
The business can now promote book sets more effectively, increasing potential sales through targeted offers and personalized marketing.
