- Chloe Crozier | cscrozi@clemson.edu
- Justin Silva | jggueva@clemson.edu
- James Tribble | jjtribb@clemson.edu
- Nayha Hussain | nayhah@clemson.edu
For Georgia Tech's Hacklytics '24 (Data Science Hackathon), we created PocDoc, an app that triages mobile health units to patients based on severity level. Patients interact with a medical-trained language model from Hugging Face to predict the severity of their condition.
We use k-means clustering, linear regression modeling (based on current local hospital burden, patient's severity level, and pre-existing conditions), and the patient's geographical location to dispatch mobile health units to a given area.
The goal is to benefit overburdened healthcare facilities in metro areas and improve access to healthcare in rural areas.
We used the OpenAI API for patient interaction and the Mapbox API to display an interactive map of patients and their assigned mobile units.
