This repository develops an advanced recommendation system to enhance the e-commerce shopping experience by automating product suggestions and analyzing user preferences through machine learning techniques.
Contains notebooks with methodologies evaluated but not selected for the final implementation.
This directory houses the core notebooks essential for understanding and replicating the recommendation system:
- Language Detection:
Final_BigData_Language Detection.ipynb - Sentiment Analysis:
Final_Dask_Sentiment_Analysis-2.ipynb - Exploratory Data Analysis:
Final_EDA_BigData.ipynb - Recommendation System:
Final_RecommenderSystem_SVD.ipynb
Utilizes Python and Jupyter Notebook, with libraries including Pandas, Numpy, Matplotlib, Seaborn, Scikit-learn, and Dask.
To set up this project locally, run the following commands:
git clone https://github.com/yourusername/your-repository-name.git
cd your-repository-name
pip install -r requirements.txt
Prepare the dataset:
- Download the necessary data files. https://www.kaggle.com/datasets/cynthiarempel/amazon-us-customer-reviews-dataset
- Place them into the archive/ folder on your Google Drive Home directory
Navigate to the final_implementation directory:
cd final_implementation- Data Visualization and EDA: Open and run
Final_EDA_BigData.ipynbto visualize and explore the data. - Language Detection: Continue with
Final_BigData_Language Detection.ipynbto detect the language of the reviews and you will have an output file in Big Data Project/ named language.parquet/ - Sentiment Analysis: Proceed with
Final_Dask_Sentiment_Analysis-2.ipynbfor analyzing sentiments using Dask and you will have an output file in Big Data Project/ named senti.parquet - Recommendation System: Finish by running
Final_RecommenderSystem_SVD.ipynbto see how the SVD-based recommender system performs.