PickSmith is a content-based movie recommendation system that implements TF-IDF vectorisation and cosine similarity. The Python engine processes plot descriptions, genre metadata, and director information to generate personalised suggestions. It is designed as a modular application that demonstrates production-ready ML pipelines for educational and practical implementations.
- Analyses movie plot descriptions using TF-IDF vectorisation text processing
- Combines genre, director, and rating data for hybrid and reliable recommendations
- Measures movie cosine similarity based on processed feature vectors
- Easy to extend with new algorithms or data sources
| File | Description |
|---|---|
app.py |
Main CLI interface for user interaction |
preprocessing.py |
Data loading and feature engineering |
rec.py |
Recommendation algorithms and scoring logic |
movies.csv |
Dataset containing movie metadata (title, plot, genres, etc.) |
test.py |
Unit tests for recommendation logic |
- Python 3
- scikit-learn (TF-IDF, Cosine similarity)
- Pandas (Data processing)
- Numpy
- Clone repo:
git clone https://github.com/aj11anuj/SeenIt.git
- Install dependencies:
pip install pandas scikit-learn numpy python
- Run:
python app.py
