France negative cases May 20

[CSSEGISandData]

Hello all,

Shortly, we will be merging in a PR that corrects the stale data we have been publishing for France since the 20th of May. On this day, France dropped nearly 350,000 cases on their dashboard. At the time, the reasoning for the drop was unclear and we wanted to see it persist in their dataset before allowing it into our data. Since then, we have become aware of this press release that explains that the reason for the drop was deduplication of individuals that had multiple positive tests. We have looked for data to correct our time series, but it does not appear to have as of yet been made available. As such, the drop in cases on May 20 will persist until we are able to find an alternative.

[yetzt]

from the information given in the press release it appears doubtful that corrected time series data will be released.

also this still does not account for the ~50000 cases discrepancy between the data published here by jhu and the official data from santé publique. (jhu: 5863138 cases, santé: 5917397 cases) — especially since santé publique is listed as the only source for french data.

[CSSEGISandData]

Hi @yetzt,

Please see #2459 (pinned) that describes the approach for France and why we need to compute the mainland France total from the total posted by the French Ministry of Health (that includes cases and deaths in Guadeloupe, Mayotte, French Guiana, Reunion, & Martinique).

[Jferrie1]

are we still pulling in this data to fix for this?