POI-RECOMMENDER

Overview

This project aims to realize a Recommender System that leverages the power of Graph Neural Networks (GNNs) to provide personalized recommendations for Points of Interest (POIs). By analyzing user preferences and historical data, the system can predict and suggest locations that a user is likely to be interested in. The core functionality revolves around a GNN architecture that captures the complex relationships and interactions between users and POIs.

During inference, users can input a query, and the system will process it alongside the user's historical data to generate tailored recommendations. This approach enhances the user experience by delivering highly relevant and personalized suggestions, making it easier for users to discover new and interesting places that align with their tastes and interests.

The project is designed to be flexible and scalable, accommodating various datasets and user inputs to provide robust and accurate recommendations.

Installation

To set up the project locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/Mahnz/poi_recommender.git
   cd poi_recommender

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install the required dependencies by executing the env_setup.py script:

   python ./lib/env_setup.py

Usage

Retrieving Venue Images

To retrieve venue images from Google Street View:

1. Set up your API credentials in the environment variables STREET_VIEW_KEY and STREET_VIEW_SECRET.
2. Run the image retrieval script:

   python preprocessing/retrieve_venue_images.py

Generating Descriptions

To generate descriptions for venue images:

1. Place the images in the images/ directory.
2. Run the description generation script:

   python preprocessing/descriptions_generator.py

Encoding Descriptions

To encode the generated descriptions:

1. Ensure the descriptions are stored in additional_data/venues_desc.csv.
2. Run the encoding script:

   python preprocessing/description_encoder.py

Acknowledgements

• PyTorch
• Hugging Face
• Google Street View API