Include hospitalization data in Marin data scraper

[Mr0grog]

In #167 I rewrote the Marin data scraper to use the new data the county is now surfacing in its Socrata data portal. One of the things the new data includes is a timeseries and demographics for hospitalizations. We don’t currently have this for other counties, so I didn’t include it because the update was a hotfix to repair broken functionality. However, now that the hotfix has shipped, now would be a good time to consider whether and how to include the new data.

No Few other counties currently surface either the timeseries of hospitalizations or the demographics (at least I don’t recall that they do), but it would still be nice to include more data if it’s available from a given county. (Update: apparently Napa has even more detailed data!)

• Should these be included? The timeseries seems at least a little duplicative of the data we are already gathering from the state. However, I don’t think the demographics are not available from the state. So maybe worth including them but not the timeseries?

• How should it be formatted? I’m thinking the timeseries (if included) would be series.hospitalizations should be an array formatted similarly to our other timeseries:

  {
    "date": "2020-05-05",
    "patients": 2,
    "cumul_patients": 37
  }

  And the demographics would be hospitalization_totals, which would be an object formatted just like case_totals.

Also, as @benghancock noted (#167 (comment)), we should probably cross-check this with the hospitalization data we are already pulling from the state to make sure they at least kinda-sorta line up.

---

Updated 2021-01-11: @1ec5 noted that Napa Co. has detailed hospitalization data, too. Nobody’s written a scraper for it yet.