An interactive Streamlit app to explore, cluster, and generate smart recommendations on Amazon products.
Includes preprocessing, KMeans clustering, visualizations (boxplots, t-SNE), and a data-driven suggestion system for sellers.
amazon_insights_app/ ├── dataset/ │ ├── amazon.csv # Original raw dataset │ ├── amazon_clean.csv # Cleaned dataset │ └── amazon_clustered.csv # Clustered dataset ├── models/ │ └── kmeans_model.pkl # Trained KMeans model ├── notebooks/ │ ├── preprocessing.ipynb # Data cleaning and transformation │ └── cluster.ipynb # Clustering analysis and t-SNE ├── pages/ │ ├── 1_Descrizione.py # Data overview and stats │ ├── 2_Clustering.py # Cluster visualizations │ └── 3_Raccomandazioni.py # Interactive recommendation system ├── app.py # Main Streamlit app entry point └── requirements.txt
The dataset contains real Amazon product information:
- Product name, price (actual/discounted), discount percentage
- Ratings and number of reviews
- Product category and description
- Image and purchase links
- 🧾 Data Overview: Preview data, inspect basic stats, plot price distributions.
- 🤖 Clustering: Group products by pricing, rating and discount using KMeans.
- 📉 Visualizations: Boxplots, category counts, t-SNE plots for dimensionality reduction.
- 📦 Recommendation Engine: Suggest the best category and product example based on user input.
- Clone the repo:
git clone https://github.com/your_username/amazon_insights_app.git cd amazon_insights_app
Install dependencies: pip install -r requirements.txt
Run the app: streamlit run app.py
🧠 Built With Python 🐍 Streamlit – for building the interactive UI Pandas & NumPy – for data wrangling Scikit-Learn – for clustering and preprocessing Matplotlib & Seaborn – for beautiful charts
👤 Author Developed with ❤️ by Dante Trabassi Feel free to open an issue or contribute via pull request!
License MIT