🧴 Sephora Skincare Recommender

📌 Overview

A Hybrid Recommendation System for Sephora skincare products, combining:

1️⃣ Item-Based Collaborative Filtering – Recommends products based on user ratings.
2️⃣ Content-Based Filtering – Recommends products based on ingredients, categories, and highlights.

🎯 Purpose

• Help users find skincare products tailored to their preferences, skin type, and concerns.
• Provide personalized recommendations using machine learning & NLP.

🚀 Features

✅ Hybrid Recommendation Model (Item-Based CF + Content-Based Filtering)
✅ Flask API – Serves recommendations dynamically
✅ Streamlit Web App – User-friendly interface to explore recommendations

📦 Structure

├── 📂 data/                     # Contains raw dataset files (ignored in Git)
│   ├── product_info.csv         # Product metadata (ID, name, brand, ingredients, categories, price, etc.)
│   ├── reviews_0-250.csv        # User reviews (ratings, skin type, feedback, etc.)
│   ├── ...                      # Other review datasets (split for large-scale handling)
│
├── 📂 metadata/                  # Stores processed data files
│   ├── products.csv              # Cleaned & processed product metadata
│   ├── hybrid_similarity.pkl     # Precomputed similarity matrix for recommendations
│
├── 📂 skincare/                  # Virtual environment (ignored in Git)
│
├── api.py                        # Flask API for serving recommendations
├── app.py                        # Streamlit Web App interface
├── model.ipynb                   # Jupyter Notebook for training & evaluating models
├── requirements.txt               # Dependencies for installing the project
├── .gitignore                     # Excludes unnecessary files (datasets, env, cache)
├── README.md                      # Project documentation (this file)

📂 Dataset

• Product Data: Contains product ID, brand, ingredients, categories, price, and highlights.
• Reviews Data: User ratings, reviews, and skincare attributes (skin type, concerns).

1️⃣ Download the Dataset

1. Go to the Sephora Skincare Reviews Dataset on Kaggle
2. Click Download and extract the ZIP file.

2️⃣ Move Files to data/ Directory

• After extraction, place the dataset files inside the data/ folder:

├── data/
│   ├── product_info.csv
│   ├── reviews_0-250.csv
│   ├── reviews_250-500.csv
│   ├── reviews_500-750.csv
│   ├── reviews_750-1250.csv
│   ├── reviews_1250-end.csv

🛠️ Installation & Setup

1️⃣ Install Dependencies

pip install -r requirements.txt

2️⃣ Run the Flask API

python api.py

• API will be available at: http://127.0.0.1:5000/recommend?product_id=P12345&num_recommendations=5

3️⃣ Run the Streamlit Web App

streamlit run app.py

• Open in browser: http://localhost:8501

📌 How It Works

1. Users select a Sephora skincare product from the web app.
2. The model generates recommendations using a combination of product similarity (content-based) and user ratings (collaborative filtering).
3. Users receive recommendations with product details (brand, price, and rating).

🛠️ Future Improvements

• Improve recommendations using sentiment analysis on user reviews.
• Personalize results further based on skin type & concerns.
• Deploy API & Web App on AWS/Heroku for public access.