Anomalies in people vaccinated corresponding to join with "high priority cdc" data

[free-soellingeraj]

I am looking at the dataset for US States and finding that there are anomalies in which the total people vaccinates goes down day-to-day. For example, 2021-02-21 in the District of Columbia has a situation where the number of people_vaccinated goes from 109,045 (15.5% of 2019 census) to 71,399 (10.1% of 2019 census). I looked at the commits and it looks like a special commit was merged in marked as the cdc. I am in the process of identifying all places where people_vaccinated goes down day-to-day.

[free-soellingeraj]

Here is the list of all dates/states on which the people_vaccinated variable goes non-monotic:

{'Arizona': [Timestamp('2021-02-22 00:00:00')],
 'Colorado': [Timestamp('2021-01-23 00:00:00')],
 'Connecticut': [Timestamp('2021-03-13 00:00:00')],
 'Delaware': [Timestamp('2021-02-22 00:00:00')],
 'District of Columbia': [Timestamp('2021-02-22 00:00:00')],
 'Florida': [Timestamp('2021-01-25 00:00:00'),
  Timestamp('2021-02-22 00:00:00')],
 'Georgia': [Timestamp('2021-02-22 00:00:00')],
 'Hawaii': [Timestamp('2021-01-23 00:00:00')],
 'Iowa': [Timestamp('2021-01-14 00:00:00')],
 'Louisiana': [Timestamp('2021-02-22 00:00:00')],
 'Missouri': [Timestamp('2021-02-22 00:00:00')],
 'New Jersey': [Timestamp('2021-01-14 00:00:00')],
 'North Carolina': [Timestamp('2021-02-22 00:00:00')],
 'North Dakota': [Timestamp('2021-02-22 00:00:00')],
 'Tennessee': [Timestamp('2021-02-22 00:00:00')],
 'Texas': [Timestamp('2021-02-25 00:00:00'), Timestamp('2021-03-24 00:00:00')],
 'Utah': [Timestamp('2021-02-22 00:00:00')],
 'West Virginia': [Timestamp('2021-02-22 00:00:00')]}

If possible, please advise on the meaning of this? For example, should I nullify all observations prior to these dates as they indicate a correction?

[edomt]

Hi @free-soellingeraj See #698