Individualized Healthcare App
Creating apps for clinicians to discover trends in individualized medicine, making AI algorithms accessible.
Training time down from 3 months to 3 days
GNS Healthcare has artificial intelligence algorithms to discover the connection between patient traits and improved outcomes for medical treatments. I worked with their team of engineers to create interactive apps to allow users to create and run models without knowing how to use R or use the command line. The training time to use these models went from 3 months down to 3 days because of the apps.
I was first given a rough draft of one app in order to redesign it. I overhauled the look in order to make it look clean, modern and in-line with the existing company branding. I used R shiny and custom CSS to achieve this. The left image is the old app and the right is the redesign I made.
I then joined the engineering team to expand on the code to add features. I completed hundreds of tasks in order to add interactive tables to display plots, inputs to allow users to group variables using regular expressions, and a network diagram so users could edit connections.
The engineers and I tracked our progress using Github. This app was made into a proprietary R package complete with documentation and tests. It is now a complete GNS Healthcare product.
I worked with a designer on a second app that allowed users to access the proformance of the models they built with the previous apps. I coded all the features and used CSS to get the app to get as close to the mockups as possible. The top image is the mockup and the bottom is the final product.
