Louisiana

State or US: LA

Describe the problem LA's total tests are stored in people, but we had made that classification based on information that appeared to only pertain to LTC. LA's language has always been in terms of "tests" and the numbers seem to indicate specimens compared to HHS data, and based on test positivity. We are going to move this to specimens, consistent with the way we've treated other states that use similar language.

Link to data source Our own people timeseries

State or US: Louisiana

Describe the problem As of 8/26, Louisiana is reporting probable cases on their dashboard. CTP will start capturing them today.

1. Backfill the values for 8/26 and 8/27
2. LA's data dictionary now confirms explicitly that reported cases are lab confirmed via a PCR test, so we should copy the cases numbers we captured previously into the Positive Cases (PCR) column
3. Add the probable values to the confirmed values in our Total cases (positive) metric.

Link to data source

State or US: Georgia, Louisiana

Describe the problem GA has -74,189 new tests on May 27. All quantity of tests should be a positive value. LA has -2,651 new test on April 19. Negative values have caused the 7-day average to scale to 300% make the rest of the 7-day average a flat line at the bottom.

Link to data source https://coronavirus.jhu.edu/testing/individual-states/georgia https://covidtracking.com/data/state/georgia#historical https://covidtracking.com/data/state/louisiana#historical

Cases reported in LA were not clear as to whether they're confirmed. However, this value was reported in the confirmed column Positive Cases (PCR) on WS2 from 4/22 - 6/30. Historicals and WS2 needs to be corrected to reflect this annotation.

On 28-May both total and negative were reported as the "total" value, which increased the total reported through the API.

Total from dashboard = 355,027 Negative should be (35,5027 - 38,802) = 316,225

Links: https://covidtracking.com/api/v1/states/LA/20200528.json