Restaurant Sentiment Analysis and Image Classification

This project performs a comprehensive sentiment analysis on restaurant reviews, leveraging Hugging Face's powerful NLP models. The workflow includes the following steps:

Sentiment Analysis of Text Reviews:

We analyze customer reviews using pre-trained sentiment analysis models from Hugging Face to derive sentiment scores (positive, negative, neutral).

Correlation with Customer Ratings:

We calculate the correlation between the derived sentiment scores and the corresponding customer ratings to understand how closely the sentiment matches the given ratings.

Image Classification using Sentiment Labels:

The sentiment scores from the text reviews are then used as labels to train a deep learning model for classifying images of restaurants. This deep learning model uses the sentiment values (positive, negative, etc.) as labels, allowing for visual sentiment classification.

Key Components:

Sentiment Analysis: Utilizing Hugging Face Bert transformer for text-based sentiment analysis.

Correlation Analysis:

Investigating the relationship between review sentiment and customer ratings.

Deep Learning Model:

Training a CNN (Convolutional Neural Network) on restaurant images using the derived sentiment labels.

Dataset:

Restaurant review text and customer ratings, along with corresponding restaurant images. Google Maps Restaurant Reviews Dataset: 'https://www.kaggle.com/datasets/denizbilginn/google-maps-restaurant-reviews'