Covid has radically altered our lives like never before. Knowing where we are on the contagion curve will help people know when this nightmare will end.
What it does
Using an ARIMA model, the app will generate predictions for the number of Covid infections.
How I built it
Data from Johns Hopkins was fitted using an ARIMA model to generate forecasts. The forecasting models are running in R while the web interface uses Shiny.
Challenges I ran into
Finding a web interface to push forecasts required some work.
Accomplishments that I'm proud of
I am glad that I was able to build an app that is simple enough for a lay person to use.
What I learned
Building a forecasting model is one thing, but putting it out as a data product is a whole different challenge.
What's next for Predicting Spread of Covid
I need to work on reducing the prediction error for this model. I am also planning on forecasting other metrics such as per capita infections or infections adjusted for population density.
Try It out