This project explores and analyzes user reviews from the Amazon Fine Food Reviews dataset. It includes two key Jupyter notebooks that work together to assess the sentiment of reviews and categorize them into meaningful product categories.
Purpose:
This notebook focuses on sentiment classification of individual words found in the reviews.
Main Features:
- Cleans and processes text data.
- Identifies and classifies frequently used words into three sentiment groups:
PositiveNegativeNeutral
- Provides an overall sentiment score based on the words' polarity.
Goal: To understand the general tone of the reviews using a word-level sentiment approach.
Purpose:
This notebook extends the sentiment analysis by mapping each review to a specific product category and aggregating sentiment at the category level.
Main Features:
- Categorizes reviews into 96 distinct product groups (e.g., Coffee, Chocolate, Dog Food, etc.).
- Aggregates overall sentiment for each product category.
- Helps identify which categories are most positively or negatively received by consumers.
Goal: To provide insights on customer sentiment across different food product types on Amazon.
- Python (Pandas, NLTK, Scikit-learn)
- Jupyter Notebook
- Text preprocessing and NLP techniques
- Sentiment analysis using lexicon-based methods
- Amazon Fine Food Reviews Dataset
Contains over 500,000 food reviews from Amazon spanning several years.
- Identify customer satisfaction trends across food product categories.
- Highlight top-performing and poorly received food items.
- Support product development and marketing strategies through review insights.
- Clone the repository.
- Install the required libraries from
requirements.txt(optional). - Run the notebooks in sequence:
- Start with sentiment classification (
sentiment analysis.ipynb) - Then run product categorization (
sentiment analysis and product categorization.ipynb)
- Start with sentiment classification (