Projects

• healthSystem [hold]:
  The purpose of this project is to design and build a lightweight and easily modifiable health data system. This design will be modular to promote customization, and avoid large monolithic repositories. Intended features are to include a database schema, ETL tools, DICOM imaging database, and automation services.

  In-progress Features:
  

  • healthdbModels: Framework to construct a simple, customizable health database (HealthDB).
  • healthdbETL: ETL and Database Interaction Tools for HealthDB

  Testing:

  • healthdbModels and healthdbETL are being tested with synthetically generated patient data using the Synthea package.

  Planned Features:

  • Lightweight Orthanc Dicom Server for advanced image storage and sharing. This service will run in a Docker container and tie into the HealthDB model.
  • Incorporation of HL7 and FHIR standards for interoperability and data exchange. (HL7apy, FHIRclient)
  • Task automation using Apache Airflow

  The project stack includes: Python, PostgreSQL

• PsychPredict [hold]:
  This project was designed to predict incident psychiatric hospitalizations using data sourced from an electronic medical record at Weill Cornell Medicine. The project incorporates SQL queries, transformations, and common machine-learning algorithms.

  The project stack includes: Python, PostgreSQL