Open MelbourneDeveloper opened 4 years ago
@MelbourneDeveloper Thanks for providing this list! Yes, there are some strange discrepancies. However, all the states 'Unassigned' decreases are easily explained by those cases being assigned to their proper counties.
@Aureuum thanks for letting me know. Is there an issue I can look at to see what that means? I can fix this issue in my database tool so that the database has the correct values.
@Aureuum can "Unassigned" locations simply be ignored?
Or is there a plan to revise the old data to fix the issues?
@MelbourneDeveloper About the 'Unassigned' cases I just include them in my presentation as they are, being both positive and negative daily changes over time. And they are usually much fewer than the assigned cases so they have less general significance. For the location of the 'Unassigned' cases, I use the same as for the state itself. I'm not aware of any plan to fix the issues.
@Aureuum
Should unassigned counts be included with other locations for a given Province / Region? Or, do the counts overlap? This is important because if they are included in a Sum of all the values for a Province / Region, there will be inaccuracies.
Is there an issue somewhere that explains what the meaning of unassigned is?
@MelbourneDeveloper This is how I interpret it: County A: 100 assigned cases County B: 200 assigned cases State unassigned cases: 50 State total cases: 100+200+50=350 cases Does that make sense? I haven't seen or looked for any definition, just assumed that it works that way.
@Aureuum that's how I originally interpreted it. How can we confirm that this is the case?
Does that mean that in cases where the unassigned drops, those cases are moved to the actual location? If that's the case, it's very hard to calculate accurate data over time for locations.
@MelbourneDeveloper To confirm that's true one would need to ask the data provider or find info about it from them. And yes, if 10 cases are moved from 'Unassigned' to its proper location, the 'Unassigned' will drop by that amount and the location increases by the same. So one would need to accept that as a factor for a location's statistics. However, in most practical cases I think it's a minor factor compared to the amount of new cases that are properly assigned every day. At the same time it's a valid factor that a location's amount of cases changes due to reallocation.
Note: Some Urls may be out of date because some rows are being deleted or moved. Some data will be fixed by the time this is viewed. Please pay attention to the Region/Province/Location and the Date.
This is a comprehensive list of all the rows where a count value has decreased by 10 or more from the previous day. There is a Url for the row, and URL for the previous date so you can clearly see the discrepancy. The discrepancy column is how much the value decreased by from the previous day.
Column
is either "New Cases", "Deaths" or "Recoveries". This list is sorted by the largest discrepancy to the smallest.This list was generated with this tool. Please contact me for help with automating data validation @CSSEGISandData