SK case data should be updated with "previous week data" in weekly report

[jeanpaulrsoucy]

The SK weekly case report contains a column with the current week's case counts as well as the previous week's case counts.

Often, data reported for the previous week is slightly different from the data reported reported for the same period in the previous week's report. For example, today's report gives 392 cases for last week, whereas last week 394 cases were reported. These are clearly the result of minor cleanup/fixes to the underlying data.

Since both the current and previous week's columns are transcribed in covid19_reports, we should take advantage of these more accurate retrospective data. This would require a more complex process for SK cases in Covid19CanadaETL.

[jeanpaulrsoucy]

Note that the first monthly report may have some "previous 7 days" data for the final weekly report.

[jeanpaulrsoucy]

I've added all of the historical hospitalization data from the reports, but the historical case data still haven't been added.

[jeanpaulrsoucy]

The historical hospitalization data don't seem to be strictly necessary, since they only seem to disagree with the previously reported values in two cases:

The first week (disagreeing with the time series value for 2022-02-02 of 372):

I will prioritize the daily time series value here, so no change.

And report from 2022-03-17:

This contradicts the previous week's 339. However, this is clearly a typo: the "difference from last week: -40" supports the original value.

The ICU data have no contradictions, not even to the time series value in the first week.

[jeanpaulrsoucy]

Should probably repeat this for the new admissions stat.