search

CHAD-Analytics / CHAD

COVID-19 Health Assessment Dashboard
MIT License
13 stars 5 forks source link

Scaling Methods #56

Open levilott opened 4 years ago

levilott commented 4 years ago

Is your feature request related to a problem? Please describe.

Currently I think both CHAD and EXAIL are scaling state level projections to the county level by the % county population to the state: County_Projection = State_Projection * (County_Population / State_Population)

Describe the solution you'd like

I think an alternative method that could be more accurate would be to scale by % county cumulative case and deaths to the state: County_Case_Projection = State_Case_Projection * (County_Cumulative_Cases / State_Cumulative_Cases) and same for deaths

This would require capturing the most recent confirmed cases and deaths and the scaling factor would be sort of dynamic with respect to the date but since were talking cumulative up to this point a small change in the number of cases and deaths wouldn't change the percentage greatly.

I think this would be a more accurate way to allocate cases and deaths to counties. I'm going to try it out on my end. Let me know what you guys think of the idea.

tleonard6 commented 4 years ago

@levilott i think this is a great idea. We would be happy to put it in place in CHAD, possibly as a pair of option.

galarcon0308 commented 4 years ago

Concur, the initial scaling method was a quick mock-up. Surprised it honestly lasted this long. But you make a great point. Cases per county gives a better look at covid probability by county, which will def be better. Excellent idea and thanks for posting!

levilott commented 4 years ago

Cool I appreciate the feedback! I hit some hiccups with IHME data that I had to tackle today but I'm getting back on this scaling method tomorrow and monday. Will update once complete!