BookRecommender

Book Recommender: Find the Book that Fits You

Author

Developed by Sunwoo Baek, Hangoo Kang, Supia Park, Esther Hwang

---

Introduction

This project processes tweets and book datasets to:

• Impute missing categories for books using zero-shot classification.
• Preprocess tweet content by removing links, mentions, and emojis.
• Classify user tweets into predefined book categories.
• Recommend books to users based on tweet analysis.

The script uses tweepy, transformers, and scikit-learn, among other libraries. It leverages Kaggle datasets and provides outputs in CSV files.

---

Requirements

Tools and Libraries

• Python 3.8+
• Kaggle API configured with your account
• Libraries: tweepy, torch, transformers, kagglehub, sklearn, pandas, csv, os, re

---

Setup Instructions

1. Install Python and Required Libraries

Ensure Python is installed on your machine. Install the necessary libraries by running:

pip install tweepy torch transformers kagglehub scikit-learn pandas

2. Download the Dataset from Kaggle or from Zip.file

This script requires tweet and book datasets from Kaggle. You must have Kaggle CLI configured on your system. Or, you can directly load datasets included in the Zip.file.

Steps to Download the Dataset:

1. Install Kaggle CLI:

   pip install kaggle

2. Download the datasets by running the following in the terminal:

   kaggle datasets download mmmarchetti/tweets-dataset

   The dataset will be downloaded as a ZIP file.

3. Extract the ZIP file:

   unzip tweets-dataset.zip -d tweets_dataset

or

1. Load books.csv and tweets.csv from the Zip.file.

---

How to Run

Step 1: Setup Environment

Ensure all required libraries are installed and datasets are unzipped and placed in the correct directories.

Step 2: Execute the Script

Run the script using Python:

python analysis_script.py

Step 3: Outputs

1. Books with Imputed Categories:

   • Saved as books_with_imputed_categories.csv.
   • Contains missing book categories imputed using zero-shot classification.

2. Preprocessed Tweets:

   • Saved as tweets_with_imputed_categories.csv.
   • Contains tweets with links, mentions, and emojis removed.

3. User Classification and Recommendations:

   • Saved as author_classification_results.csv.
   • Contains the predicted top category for each author and the top 5 recommended books.

---

File Structure

project/
├── books.csv                         # Dataset with book information
├── tweets.csv                        # Dataset with Tweets
├── books_with_imputed_categories.csv # Output with imputed book categories
├── tweets_with_imputed_categories.csv# Preprocessed tweet data
├── author_classification_results.csv # User classification results
├── CS410_Project_BookRecommender.ipynb           # Main Project Jupyter Notebook

---

Notes

• For failed classifications or missing data, a fallback category (e.g., Fiction) is used.
• Use the Kaggle API token in your system to download datasets programmatically or use datasets that are included in the zip file.

---

Troubleshooting

1. Missing Datasets:
   • Ensure the datasets are downloaded and unzipped in the correct directory.
2. Dependency Errors:
   • Check Python version and installed libraries.

---

License

This project is for educational purposes only. Use of datasets and models must comply with their respective licenses.

---