Scribe AI
Problem Statement
- Retrieving critical and relevant data from Electronic Health Records can be an incredibly arduous task containing decades of encounters and multiple sources of information.
- Reduce friction between physicians and EHR software.
Solution
-
Full-stack software that takes physician input and uses a Machine Learning model to perform a classification task.
-
Relevant information is pulled from the database and returned to the physician.
-
Frontend
- React.js, Socket.io
-
Backend
- Django, MongoDB, Azure, PyMongo
-
ML Model
- Spacy (SVM)
- NLP Text Categorizer
Start Backend
- install all requred libraries
- cd django-react-scribeai/backend
- daphne backend.asgi:application -p 8000
Start Frontend
- cd into /django/frontend
- npm i
- run npm start