Project name: Sentiment Analysis on Product Reviews

Project description

This project involves developing a Python program for sentiment analysis on a dataset of product reviews. The program, named sentiment_analysis.py, utilizes the spaCy library for natural language processing tasks to analyze and classify the sentiment of the reviews. Additional codes - testing_exception_stop_words.py and testing_exception_stop_words.py - test the validity of the used sentiment method from SpacyTextBlob.

Table of Contents

1. Installation
2. Usage
3. Credits

Installation

To install and run this project locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/jsymonowicz/ML--Review-Sentiment-Analysis-Model
   

2. Install dependencies:

   • Download the spaCy English model:

     python -m spacy download en_core_web_sm
     

   • Install the spacytextblob package:

     pip install spacytextblob
     

   • Download TextBlob corpora:

     python -m textblob.download_corpora
     

3. Download the dataset: Download the Consumer Reviews of Amazon Products dataset named amazon_product_reviews.csv (Source: https://www.kaggle.com/datasets/datafiniti/consumer-reviews-of-amazon-products, "1429_1.csv")

Usage

Read instructions and comments in 'sentiment_analysis_report.pdf'. Execute codes in the Python IDE of your choosing (e.g. Spyder). The order does not matter.

• sentiment_analysis.py: conducts sentiment analysis on a dataset of product reviews and also performs similarity analysis.

• testing_text_preprocessing.py: evaluates the impact of removing stop words, punctuation, and spaces on the polarity score of a review. The goal is to determine the most suitable text preprocessing methods for preserving the true sentiment.

• testing_exception_stop_words.py: script dedicated to testing which stop words should not be removed from the predict_sentiment function to maintain the accuracy of sentiment analysis.

Credits

This project was developed by Joanna Symonowicz as part of HyperionDev Data Science BootCamp.

If you find this project useful, consider giving it a star on GitHub and citing the authors. Contributions and feedback are welcome!