It seems like the HN community is well positioned to actually do this. So I could host it but I'd need others to help build it and source data.
Edit. I just registered c19testlocations.com
https://news.ycombinator.com/item?id=22577611
But will support any effort so long as we get it done. I can help from the medical and the technical side of things. What’s the best way to sync up?
Lot of test are useful if you can vaccine all the people that meet the (false) positives cases, and all the people that meet the people that meet the (false) positives. This is a vaccination rings https://en.wikipedia.org/wiki/Ring_vaccination but there are no vaccines for now.
Another possibility is to quarantine all that people for two weeks just in case, but if someone disagree you must convince the police to enforce that, and hopefully the police is following the advice of the CDC and want an official test.
To get something up and running, it could be worth searching the web for services and code that solve similar problems and starting with a proof-of-concept or draft version.
The solution sounds similar to a 'store locator' (like USPS / large retail chains provide); using that in some searches yielded the following:
- https://www.mapcustomizer.com - this looks basic but allows plotting locations on a map and provides bulk entry, which could be performed and managed from a Google Sheet by a team
- https://batchgeo.com/features/store-locator/ - these folks specifically offer a service for building a store locator and they also provide geocoding (translating addresses -> latitude and longitude). It's not free but looks inexpensive -- at least for a small number of user accounts.
- [more advanced, would require self-hosting] Google have a tutorial on creating your own store locator using PHP and MySQL with Google Maps here: https://developers.google.com/maps/solutions/store-locator/c...
- [most advanced, caveats apply] if building something from scratch, Solr geospatial[0] and leaflet.js could be starting points, along with something like uszipcode[1] to determine the user's origin long/lat to search-by-distance. Ultimately an approach like this risks sinking time into developing scripts and data processing that replicate what others already provide online as a service. It would provide more control over the experience and functionality however.
The concerns other commenters raise about false positives are valid - but it's likely that if such a service does become successful enough to produce noticeable test center foot traffic, the development team could communicate with those health services to update the messaging as-needed (or pause the service).
I'd encourage posting any data you collect and geocode to an open source repository in GitHub or elsewhere so that others could build related tools and/or contribute improvements and corrections back.
As an example of that approach: Johns Hopkins have been great at sharing the C19 global case data they're gathering[2] and at the time of writing there are 3.2k forks created from that dataset which is a good initial indicator of interest.
Good luck!
[0] - https://cwiki.apache.org/confluence/display/SOLR/SpatialSear...