HealthCruncher

Live demo

athenahealth hackathon 2015

Netflix but for predicting customers' health outcomes.
💵 "B2B SaaS" (Business to business software as a service)

It is also an API that (will) attempt to follow the principles of REST 😴

Development

• Some stuff we use: 🐍 flask, DigitalOcean, scikit-learn, anaconda
• $ python main.py
• Test at 0.0.0.0:5000
• Use a virtualenv if you feel like it.

Deployment

• $ ssh root@strtup.me
• $ cd /var/www/healthcruncher
• $ git pull
• In main.py, debug should be FALSE!!!
• $ . venv/bin/activate
• 🐍 $ nohup python main.py > /dev/null& (TBD)

Undeployment

• $ ps -fA | grep python
• Find the <PID> of the one that has python main.py
• kill <PID>
• Save some 💵 , shut it down

Todo

• HTTPS
• Better the algorithm (actual data, better validation)
• Better the API (fix some Flask routing issues)
• Better form validation (CSRF protection!)
• Streamline deployment (scripts, etc.)

Primary roles

Rainier Ababao - web dev/deployment/data science
Shaun Mataire - iOS/deployment
Hayley Call - web dev/business model
Sarah Gorring - web dev/business model

Our original idea

A web app that accepts user input on easily accessible patient data and output metrics based on a machine learning algorithm.