How chhssmch/movie-recsys works
This is an educational MovieLens recommendation system. Its main workflow loads and prepares ratings and movie metadata, builds four recommendation approaches, evaluates their ranked results, and saves metrics. The graph also includes the standalone popularity/genre and similar-item capabilities. The README describes production concerns and monitoring as coursework, not implemented runtime services, so those are not shown as system components.
Main entry points: data_loader.py Data loading and splitting · evaluate.py Model evaluation · run_all.py Pipeline runner · popularity.py Popularity and genre ranking
Components 11
Data preparation
- MovieLens raw data
data/raw - Data loading and splitting
src/data_loader.py - Processed datasets
data/processed
Recommendation models
- Popularity and genre ranking
src/popularity.py - Content-based recommendations
src/content_based.py - Item-based CF
src/item_based_cf.py - User-based CF
src/user_based_cf.py
Evaluation and results
- Model evaluation
src/evaluate.py - Evaluation metrics
reports/metrics.csv
Workflow entry
- Pipeline runner
scripts/run_all.py
Other components
- Analyst
Connections 16
- Analyst → Pipeline runner: starts pipeline
README.md - Pipeline runner → Data loading and splitting: loads and splits
scripts/run_all.py - Data loading and splitting → MovieLens raw data: reads
src/data_loader.py - Pipeline runner → Data loading and splitting: saves processed data
scripts/run_all.py - Data loading and splitting → Processed datasets: writes
src/data_loader.py - Pipeline runner → Model evaluation: evaluates models
scripts/run_all.py - Model evaluation → Data loading and splitting: loads and splits
src/evaluate.py - Model evaluation → Popularity and genre ranking: builds baseline
README.md - Model evaluation → Content-based recommendations: builds recommender
src/evaluate.py - Model evaluation → Item-based CF: builds recommender
src/evaluate.py - Model evaluation → User-based CF: builds recommender
src/evaluate.py - Content-based recommendations → Data loading and splitting: loads movie metadata
src/content_based.py - Item-based CF → Data loading and splitting: loads ratings and movies
src/item_based_cf.py - User-based CF → Data loading and splitting: loads ratings and movies
src/user_based_cf.py - Popularity and genre ranking → Data loading and splitting: loads ratings and movies
src/popularity.py - Pipeline runner → Evaluation metrics: writes results
scripts/run_all.py