README.md

In hoir/covid-data: What the Package Does (One Line, Title Case)

COVID dashboard data for Cambridge area

Data repository for COVID-19 data in Cambridge, MA area. Raw files from original sources (listed below) are in the data-raw/ directory. Scripts and processed data in the roadmap.

Updating data

The process is automated-ish using R drake.

To install dependencies, run renv::restore().

To update data, run r_make(). Actually need to double-check that this will download the latest files...

Processed data

Processed data is in the data/ directory. Preliminary plan is to follow the structure of the covidtracking API.

Calculated variables:

• Diffs/increases for more variables, in particular confirmed cases/deaths
• 7-day averages for a number of metrics
• Suffixed with 7d, are all 7-day means (multiply by 7 for weekly counts when desired)
• Rate of change of 7d averages, named d[OriginalVariable]7d

Data sources

Update: After looking into this a little more, https://covidtracking.com is probably the best source for US and state data. They have an API, and for MA the numbers from the daily dashboard match. NB - positivity rate does not appear to be calculable using the Covid Tracking data; - also, the MA daily dashboard reports new confirmed cases, which would need to be constructed by diffing the cumulatives.

Cambridge:

• Cumulative cases by date
• Confirmed cases in Cambridge
• Need to scrape daily to get time series
• Includes breakdowns by gender, age
• Cambridge COVID-19 dashboard
• Includes some data not in API, e.g., cases by neighborhood, but unclear how it could be retrieved

Boston (no data in repo):

Doesn't look like there's easy automated access, especially for historical data.

• bhpc.org main COVID-19 page
• Boston dashboard

Massachusetts:

• Mass.gov COVID response reporting
• Raw daily dashboard available daily; either scrape link or construct URL (seems consistent)
• Weekly public health report contains case counts by town
  • Link to raw data for latest weekly report
  • Prior weekly report raw data seems accessible by manually constructing URL

US (CDC API):

• Case surveillance public use data not in repo
• COVID-19 cases/deaths by state
• Can use API params to aggregate: https://data.cdc.gov/resource/9mfq-cb36.json?%24select=submission_date%2C+sum(tot_cases)+AS+tot_cases%2C+sum(new_case)+AS+new_case%2C+sum(tot_death)+AS+tot_death%2C+sum(new_death)+AS+new_death&%24group=submission_date&%24order=submission_date

Other data sources:

Github repos:

• NYTimes: https://github.com/nytimes/covid-19-data
• JHU
• Covidtracking: https://api.covidtracking.com
• CovidActNow.org: https://data.covidactnow.org

hoir/covid-data documentation built on Sept. 27, 2021, 10:31 p.m.