Market Basket Analysis in R

A retail grocery store needed to understand customer purchasing patterns and product associations to optimize their store layout and boost sales. The analysis required processing transaction data to identify frequently purchased items and discover meaningful relationships between products that could inform merchandising decisions.

The analysis utilized the following strategy:

• Process and clean transaction data from groceries.csv
• Convert raw data into a transaction format suitable for association analysis
• Identify top-selling products and their contribution to overall sales
• Generate association rules using the Apriori algorithm
• Visualize product relationships through interactive graphs

The implementation involved several key steps using R:

1. Transformed raw CSV data into a transaction database using the arules package
2. Applied the Apriori algorithm with carefully tuned parameters (support = 0.001, confidence = 0.80)
3. Created frequency plots to visualize top 20 selling items
4. Generated association rules to understand product relationships
5. Visualized rules using arulesViz package to create an interactive network graph

The analysis revealed several significant findings:

• The top 20 items accounted for approximately 50.37% of total sales
• Strong associations were discovered between certain product categories, particularly beverages and fresh produce
• Based on the findings, three key recommendations were developed:
  1. Implement promotional offers on alcoholic beverages
  2. Create dedicated aisles combining vegetables, fruits, herbs, rice, juices, and meat
  3. Position vegetables and whole milk centrally, surrounded by complementary products