Amazon Co-Purchasing Network Analysis
This project explores the Amazon product co-purchasing network using graph theory. By applying Tarjan's algorithm to detect communities and measuring Jaccard similarity, I analyzed how products are interconnected through customer purchasing behavior.
-
Tarjan’s Strongly Connected Components (SCC)
Identified product communities within the network
Used depth-first search and stack structure via Rust’s petgraph crate
-
Jaccard Similarity
Measured product similarity based on common neighbors
Helped determine closely related products in each community
-
Common Customer Count
Analyzed how often product pairs shared buyers
Sorted product pairs by shared customer frequency
-
Language: Rust
Used petgraph, HashMap, and custom graph functions
Focused on performance and memory efficiency