An NLP-based sentiment analysis tool to automatically classify customer reviews and predict star ratings.
This project addresses the challenge of manually sifting through large volumes of customer feedback. By leveraging Natural Language Processing (NLP) and Machine Learning, this tool provides a fast and scalable solution for understanding customer sentiment. The core of this project is a model that analyzes text from customer reviews and predicts a corresponding star rating, providing actionable insights into product satisfaction.
The key features of this tool include:
-
Automated Sentiment Classification: Automatically categorizes reviews as positive, neutral, or negative.
-
Star Rating Prediction: A machine learning model that predicts the star rating (1 to 5) based on the review text.
-
Scalable Solution: Designed to handle thousands of reviews efficiently, saving time and resources.
-
Actionable Insights: Enables businesses to quickly identify product strengths and weaknesses, and prioritize feedback for product development.
Source Link: https://www.kaggle.com/competitions/mlp-term-2-2025-kaggle-assignment-3/data
The model is trained on a dataset of customer reviews. The dataset should contain:
- id: The index column
- store_name: Name of the store
- category: Type of restaurant
- store_address: Address of the restaurant
- latitude: Latitude
- longitude: Longitude
- rating_count: Current count of ratings received
- review_time: Time since the posting of review
- review: review provided by the customer
- rating: Rating given by the customer (target variable)
-
Python: The primary programming language.
-
Natural Language Processing (NLP): The core technique used for text analysis.
-
Sentiment Analysis: The specific application of NLP to determine the emotional tone of text.
-
Machine Learning: Used to build the predictive model.
-
Data Visualization: To present insights and model performance (e.g., using Matplotlib).
-
Pandas & NumPy: For data manipulation and numerical operations.
-
NLTK & TextBlob: Libraries for NLP tasks like tokenization and sentiment scoring.