Wine Pairing Recommendation Interface

Project

This project is a Wine Pairing Recommendation System built with Python that provides personalized wine recommendations based on food descriptors. It leverages web scraping (Selenium), Natural Language Processing (NLP) techniques like Word2Vec, TF-IDF, PCA, and clustering (K-Means) for pairing wines with food. Aroma-based pairings are calculated using cosine similarity, and the results are visualized with Plotly, Matplotlib, and served via a Streamlit web interface.

Data

The data used in this project is obtained from several online wine shops.

Key Features

• Wine Database Construction: Built a wine database using web scraping with Selenium to collect wine details from online sources.
• Descriptor Analysis: Applied NLP techniques such as Word2Vec, TF-IDF, and PCA to analyze wine and food descriptors for better matching.
• Wine Pairing: Used K-Means clustering for grouping similar wine types. Calculated cosine similarity between wine and food descriptors to recommend suitable pairings.
• Visualization: Visualized the wine pairing results using Matplotlib (radar plots) and Plotly (interactive charts).
• Streamlit Interface: Developed a user-friendly interface with Streamlit for real-time food and wine pairing suggestions.

Technologies Used

• Python: Main programming language
• Selenium: For web scraping and building a wine database
• Word2Vec: NLP technique for word embeddings
• TF-IDF: Text feature extraction for food and wine descriptors
• PCA: Dimensionality reduction for feature analysis
• K-Means: Clustering wines for better pairing analysis
• Cosine Similarity: For measuring similarity between aromas
• Matplotlib: For radar plots and visualizations
• Plotly: For interactive plots and charts
• Streamlit: For building the web interface

Interface visualization (Streamlit)

TBC

Usage

To run this notebook, simply clone this repository and open it in Jupyter Notebook or Google Colab. You can also view it on GitHub.