Merge data from both JHU CSSE and CDS feeds about covid-19-chart

[davidbau]

I have hooked up Coronadatascraper.com as an alternate data feed, but it has some oddball issues.

https://covid19chart.org/?feed=cds

• All the numbers seem to be (much) smaller than the JHU CSSE numbers. Not sure if it's because they cut their daily numbers at a different time of day, or some other bug or problem.

When I flip to "US" mode, my code has a bug in finding the first day with data, starting at 3/10, while in Intl mode, US data is available going back to 1/22.

[davidbau]

Not all the numbers. Many numbers are identical. Maybe scraping problems at CDS, so merging with CSSE data seems like a good idea.

[davidbau]

There is another promising data source from the new york times, which includes county-level information also.

https://github.com/nytimes/covid-19-data

[davidbau]

Much of the work to be done is described in issue #34

[davidbau]

CDS data quality is disappointing, so I'm going to abandon the merged_feeds branch where CDS is merged for now. The only advantage of CDS is that it has more fine-grained breakdowns of provinces in italy, germany, russia.

But the U.S. data is very inaccurate, and I have no way to verify the quality of the international data, so I don't trust it.