RentWatchAI

Overview

RentWatchAI is an NLP-driven project that leverages large language models (LLMs) and Hugging Face transformers like RoBERTa to analyze sentiment in apartment reviews across various subreddits. The goal is to help renters make informed decisions by identifying problematic apartments based on user-generated content. The project uses PRAW to scrape Reddit posts and employs transformers for sentiment scoring and text summarization.

Table of Contents

• Features
• Installation
• Usage
• Steps
• Results
• Discussion
• Contributing
• License

Features

• Scrapes Reddit posts using PRAW
• Sentiment analysis with Hugging Face transformers (e.g., RoBERTa)
• Summarizes negative apartment reviews (e.g., OpenAI)
• Displays results in a user-friendly Streamlit app with sentiment scores and charts

Installation

To get started, clone the repository and install the required libraries:

git clone https://github.com/yourusername/rentwatchai.git
cd rentwatchai
pip install transformers praw matplotlib pandas streamlit

Usage

Create a Reddit application to get your credentials (client ID, client secret, and user agent). Update the config.py file with your Reddit credentials. Run the Streamlit app:

streamlit run app.py

Steps

Step 1: Install Required Libraries

Ensure you have the necessary libraries:

pip install transformers praw matplotlib pandas streamlit

Step 2: Pull Reddit Comments

Use PRAW to pull comments from specific Reddit posts.

Step 3: Tokenize Comments

Tokenize the comments to prepare them for the RoBERTa model.

Step 4: Sentiment Analysis

Utilize the RoBERTa model to predict sentiment scores and summarize negative reviews.

Step 5: Visualize Results

The results are displayed in a user-friendly Streamlit app, providing sentiment scores and visualizations.

Results

The output includes predicted sentiment scores for each review, a summary of negative reviews, and a bar chart visualizing sentiment distribution across the dataset.

Discussion

The sentiment analysis reveals insights into user experiences related to various apartments, highlighting both positive and negative sentiments. This information can be valuable for renters and real estate professionals to understand market trends and customer feedback.

Contributing

Contributions are welcome! Please submit a pull request or open an issue to discuss any changes or improvements.

Acknowledgements

• Hugging Face for the RoBERTa model.
• PRAW for accessing Reddit data.
• Matplotlib for data visualization.
• Streamlit for creating the user interface.