Advancing to the semifinals of the Novartis-sponsored NEST 2025 Hackathon, this project uses AI to transform clinical trial design processes by providing advanced recommendations.
This repository hosts our innovative project developed for the NEST 2025 Hackathon. We use semantic technology and knowledge graphs to enhance clinical trial recommendations. Our system processes and retrieves clinical trials to offer strategic insights and effective recommendations.
Explore our Neo4j graph through an interactive HTML visualization showcasing the relationships between clinical trials.
- Install Requirements
pip install -r requirements.txt
- Run the Scripts
- Set up the database and prepare the environment by running the provided scripts in sequence:
CreateRelationship.pySimilarEntities.pyingest.pymodel.py
- Set up the database and prepare the environment by running the provided scripts in sequence:
- Data Preprocessing: Scripts to clean and prepare clinical trial data.
- Knowledge Graph Construction: Uses Neo4j to build a scalable graph representing clinical trial data.
- Recommendation Engine: Utilizes Jaccard similarity metrics within Neo4j's GDS to rank similar clinical trials.
- Semifinals in NEST Hackathon 2025: Our approach has demonstrated significant potential to enhance clinical trial designs, advancing to the semifinals.
- Effective Use of AI: Leveraging AI to derive meaningful insights from complex data.
- User Guide: Detailed instructions on setting up and using the repository.
Here are some visual highlights from various stages of the project:
This image showcases the initial data preprocessing phase where clinical trial data is cleaned and prepared for analysis.
Below is a snapshot of the output from our recommendation engine, illustrating the type of clinical trial recommendations generated by our system.
Sample output of recommended clinical trials