CHIME: COVID-19 Hospital Impact Model for Epidemics

Stage

Testing

README

Hey! If you're at this page you're hopefully looking to help out with the CHIME project. Please join us in Slack or on github to join in!

Help needed

In order to ensure the CHIME team can focus on modeling and interacting with stakeholders, we're looking for help filling the following roles:

Project / product management: help with organizing the project. Help make sure that work around explaining and improving the CHIME dashboard is well organized. Help us document the project so volunteers can contribute.
Devops: help ensure that the dashboard can handle the increase in traffic. Currently we are running on Heroku, but may switch to Linode, since they have provided free resources.
Python: help us improve, maintain, and develop the app

Project Activity
Log Buzz

Update #5

We have here a dash application embedded inside a flask app as a context, being served by node. Callbacks work, db works, login / logout works. The test server is live on heroku, and the code is hosted on git. Basic instructions for running locally are in the README. Although this is aimed at UNIX users. The next goal is to write thorough documentation on developing locally and the deployment procedures. If anyone is interested in contributing in the meantime, please join #chime-app!

Update #4

CHIME 2.0 kickoff!

Due to the model's diminishing returns as the epidemic progresses, the app is going to undergo some model and platform updates so it can continue to be a resource for decision makers!

The next iteration of CHIME development will move to

Make the model it's own package so development of the model happens independently from the platform
Move platform to tech that allows user to import csv
Update model to use better methods of estimating parameters

We're looking for python developers with data modeling (especially Bayesian methods) experience or Dash experience!

MODELING THE PANDEMIC: Penn Medicine and Code for Philly have developed a tool that predicts local hospital surges due to Covid-19. Its ultimate accuracy depends on you

Read the full article on thephiladelphiacitizen.org… Published 13 Apr 2020

Update #3

Getting a lot of requests for updating/improving the code. Need to consider more in-depth frameworks for testing python code and turning app script into modules. Experienced python developers - come thru!

Update #2

Code for Philly set up a fork of the repository today to help manage incoming user and contributor interest: https://github.com/CodeForPhilly/chime

Major efforts in the new repository included:

Setting up a documentation site and consolidating knowledge: https://codeforphilly.github.io/chime/
Developing a kubernetes deployment: https://github.com/CodeForPhilly/chime/tree/master/k8s
Deploying a pre-production version of the application to Linode's LKE managed kubernetes: http://lke.penn-chime.phl.io/

Next up

Merge improvements from pennsignal's repository
Expand documentation on how to test the Kubernetes deployment locally on minikube
Automate building container and deploying to LKE on merge to master with GitHub Action workflow
Add documentation on how local groups can set up their own production deployment

Penn Medicine released a digital tool to help hospitals with COVID-19 capacity planning

The open-source project can be adapted to fit hospital populations anywhere.

Read the full article on technical.ly… Published 17 Mar 2020

Update #1

Chime is officially launched! See this tweet, and accompanying blog post for more details. The team is now focused on two things:

implementing features requested by government and medical system stakeholders
ensuring we can scale to support the traffic coming in

Help needed

In order to ensure the CHIME team can focus on modeling and interacting with stakeholders, we're looking for help filling the following roles:

Project / product management: help with organizing the project. Help make sure that work around explaining and improving the CHIME dashboard is well organized. Help us document the project so volunteers can contribute.
Devops: help ensure that the dashboard can handle the increase in traffic. Currently we are running on Heroku, but may switch to Linode, since they have provided free resources.