This project focuses on analyzing Amazon product reviews using Natural Language Processing (NLP) techniques to classify customer sentiment and demonstrate a sentiment-aware recommendation approach. The goal is to extract actionable insights from large volumes of textual review data and support better product understanding and decision-making.
The project covers the complete data analysis lifecycle — from data cleaning and preprocessing to feature engineering, model building, and evaluation.
- Python
- Pandas, NumPy
- Scikit-learn
- Natural Language Processing (NLP)
- TF-IDF Vectorization
- Jupyter Notebook
- Amazon product review dataset
- Includes customer reviews and ratings
- Text data required extensive preprocessing to handle noise, missing values, and inconsistencies
- Removed missing and duplicate values
- Cleaned text data (lowercasing, punctuation removal)
- Tokenization and stopword removal
- Converted text into numerical features using TF-IDF
- Analyzed sentiment distribution across reviews
- Identified patterns between ratings and textual sentiment
- Gained insights into customer feedback trends
- Built machine learning models to classify review sentiment (positive / negative)
- Evaluated model performance using accuracy and classification metrics
- Achieved high model accuracy (~89%), demonstrating effective sentiment prediction
- Used predicted sentiment scores to support a basic recommendation approach
- Highlighted products with consistently positive sentiment
- Demonstrated how sentiment insights can support business decisions
- Successfully classified customer sentiment from textual reviews
- Demonstrated how NLP can convert unstructured text into structured insights
- Showed practical application of sentiment analysis in e-commerce use cases
- Clone the repository
git clone https://github.com/aishwarya031298/amazon-review-sentiment-recommender.git