search

aatishb / covidtrends

Tracking the growth of COVID-19 Cases worldwide
https://aatishb.com/covidtrends/
MIT License
301 stars 107 forks source link

Feature Suggestion: Enable USA Counties as Subset of States #44

Open sbonds opened 4 years ago

sbonds commented 4 years ago

Similar to how when the USA states appear as the selection choices when the "US" region is chosen, would it be difficult to add another level of selection for each state and have the counties appear as choices for that state? The raw data for that appears to be in the NYTimes data set.

Thanks for the excellent work on this incredibly helpful and open site.

jthayilblanchard commented 4 years ago

Would second this request. I think there are 700 separate outbreaks in the US as each metropolitan area takes different measures. I'd love to see how King County, WA (Seattle) has gone differently than NY, for example, over the last month.

sbonds commented 4 years ago

I made a mock-up of one working method here:

http://covidtrends.covid19.sbonds.org/

The UI is horrific, but it works. The code is in the linked pull request #46

sbonds commented 4 years ago

A testament to the flexibility of the methods that Aatish chose-- here's your comparison on the copy I made:

http://covidtrends.covid19.sbonds.org/?region=US+Counties&location=New+York%2FNew+York+City&location=Washington%2FKing

sbonds commented 4 years ago

Actually... use caution trusting the data my fork produces. It seems like there's a problem in there somewhere as some county totals don't add up correctly

jthayilblanchard commented 4 years ago

Wow, this is just what I was looking for. Thanks for your effort.

How hard would it be to alter the y-scale bucket, so maybe to view new infections daily, or in 3 day buckets? I’m curious whether that would add some sensitivity, as I’m guessing a weekly bucket has a 2-3 week lag time as far as actions vs outcomes? JTB

On Apr 3, 2020, at 12:30 PM, Steve Bonds notifications@github.com wrote:

A testament to the flexibility of the methods that Aatish chose-- here's your comparison on the copy I made:

http://covidtrends.covid19.sbonds.org/?region=US+Counties&location=New+York%2FNew+York+City&location=Washington%2FKing http://covidtrends.covid19.sbonds.org/?region=US+Counties&location=New+York%2FNew+York+City&location=Washington%2FKing — You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/aatishb/covidtrends/issues/44#issuecomment-608619583, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHSGN7JOT4G2IAOQRXX5UGLRKY2PHANCNFSM4L2XE7LQ.

sbonds commented 4 years ago

The bad data I was seeing was caused by a name collision when two or more states had a county of the same name. I've fixed it by indexing the data using the combined state + county instead of just the county name.

Your other suggestion is interesting but probably should be its own issue/feature suggestion rather than riding along with this one.

bsmith1180 commented 4 years ago

I took a crack at this in PR #62, which can be viewed here: https://bsmith1180.github.io/covidtrends/

aatishb commented 4 years ago

Suggesting a different organization for this: What if when US is selected in the region selector, we display an addition selector beneath it that allows you to select the state? This way we don't clutter the region selector, and the fine-grained info only becomes available to those looking for it.