• 

  Overall UI

• 

  Chatbot feature

TrustLens Health Project Writeup

Inspiration

TrustLens Health was inspired by the need to make hospital capability review more transparent. Instead of only showing a score, the project connects each hospital signal to the evidence behind it so users can understand why a facility appears strong, partial, or weak for a given capability.

What I Learned

I learned how to combine healthcare facility data, evidence extraction, geospatial context, and AI-assisted explanations into one workflow using Databrick tools. I also learned about loading, filtering, evidence review, and building chatbot interactions.

How I Built It

I built TrustLens Health with a React and TypeScript frontend, FastAPI backend, and Databricks for data ingestion and deployment. The app loads facility data from Unity Catalog, normalizes hospital records, filters noisy facility names, evaluates capabilities, and displays evidence-backed signals. I also added embedded maps, expandable evidence rows, relevance indicators, and a chatbot assistant powered by Databricks AI when configured.

Challenges

The biggest challenges were keeping the evidence review simple while still showing enough detail, making the UI readable as more features were added, and ensuring the app worked both with Databricks data and local fallback data. I also had to tune filtering, evidence highlighting, and chatbot responses so the dashboard stayed useful without overwhelming the user.

Built With

• fastapi
• python
• react
• sql-warehouse
• typescript