🧾 Project Prompt (Assignment)
Crea un sistema que recomiende elementos similares a partir de características
Your product needs recommendations. Build a simple recommender and prove it works offline.
Build a recommendation system that suggests items to a user based on past interactions.
Requirements
- Create/choose an interactions dataset: (user_id, item_id, rating or click, timestamp optional)
- Implement at least:
- Popularity-based baseline
- One personalized method (kNN CF, matrix factorization, etc.)
- Provide a function or endpoint: “recommend for user X”
- Evaluate with a basic metric (e.g., hit-rate@K, precision@K)
🛠 Tools (Tech Stack)
- Python
- pandas + numpy
- scikit-learn (optional)
- Surprise (optional)
💡 Implementation Hints
- Always compare against the baseline (popularity)
- Handle cold-start users (fallback to popular items)
- If you have timestamps, do a time-aware split
- Save processed datasets and recommendation outputs for inspection