Talk:COVID-19 pandemic in Missouri

Iron County

Iron Co. is missing from this listing of Missouri counties. Please try to include it. — Preceding unsigned comment added by 2600:1702:2590:23D0:EC01:C8BC:27A9:BDCA (talk) 19:33, 8 April 2020 (UTC)

Iron County has no confirmed cases as of Apr 10. As such, it is not included on the list. Should cases develop there, it will be added.

Representing prevalence on a map

I've made maps of most other states, but am unclear about Missouri (and skipped it for that reason). This mostly because there's an extra entry in the table for Kansas City, which is not a county. What would be the best way to represent this data geographically on a county map? Should I add up the cases of the KCMO counties and Kansas City, and add up the populations of the counties, and then calculate the percentage? Some insights from someone who knows the geographic situation better, would be appreciated - thanks! effeietsanders 07:05, 2 May 2020 (UTC)

Toggle

Not sure how to fix it, but if you click both 'last 15 days' and 'may', you actually do NOT see the days in may. I guess that the 'toggle' buttons cancel eachother out. effeietsanders 00:53, 8 May 2020 (UTC)

Missouri populations

Apologies - I am new to this - just want to draw this to someone's attention.

Many of the population figures appear to be incorrect. This is giving incorrect statistics in the last column (cases per 100k of population). Clearly St Louis county cannot have almost 50% of population infected with virus (hopefully). But not only St Louis county. There are several I saw that are significantly erroneous.

Alert Ted (talk) 12:05, 9 May 2020 (UTC)

The data is from Template:COVID-19 pandemic data/Missouri medical cases by county with population figures added by Mr Xaero. Some of are indeed way off, e.g. St. Louis County, Missouri and Scott County, Missouri. PrimeHunter (talk) 12:48, 9 May 2020 (UTC)

I agree that the population of St Louis County has not the right population in the table. Based on St. Louis County, Missouri I would expect something closer to a million. St. Louis City should have something closer to 300k. Looks like the population column is messed up (probably sorted wrong before pasting?) @Mr Xaero: awaiting a fix/checkin, I'm removing the population and prevalence numbers from the template. Feel free to put back as you're double checking the numbers. effeietsanders 18:56, 9 May 2020 (UTC)

Alert Ted and effeietsanders, the population numbers were taken from the List of counties in Missouri article which in turn came from the Census data. These numbers could be off as I input all the data for all the counties in all of the states and territories (there are over 3200 counties). Feel free to input the new values from the 2010 Census as the data for the 2020 is not yet available. As for the calculation for the Cases/100k, I used the this Excel spreadsheet formula to calculate it. =ROUND((B3/E3)*100000,1) B3 is Cases and E3 is population. — Mr Xaero ^☎️ 01:46, 10 May, 2020‎ (UTC)

I have fixed the data to show the correct populations and cases/100k based on the 2010 Census data. — Mr Xaero ^☎️ 02:18, 10 May 2020 (UTC)

Thanks, that looks about right to me. (Maybe you have some thoughts for my question above too?) effeietsanders 02:46, 10 May 2020 (UTC)

What question is that?

Are you referring to the Kansas City, MO entry?

If so, since MO-DHSS is reporting Kansas City and Joplin (as both span multiple counties), I added them to the template. However, the totals of those two cities are not added to the total population as the counties that those cities are in are included in the total population. Those two cities are not listed in the total county count either as there are only 115 counties within Missouri (this includes the City of Saint Louis as it is classified as independent). — Mr Xaero ^☎️ 03:02, 10 May 2020 (UTC)

As for the initial version of the template and the population data, I have no idea how the values were so messed up. Either I sorted wrong or my script that adds commas in the appropriate locations truncated the last portion of the number. I have no idea how that occurred. Sorry about that and thanks for catching the mistake. — Mr Xaero ^☎️ 03:28, 10 May 2020 (UTC)

Yeah, Kansas City is odd in the Johns Hopkins data. Joplin does not appear in my sheet though. For Kansas City, it's mostly unclear how the reports are defining its area. Where did you find that this spans Jackson, Platte, Cass and Clay (and not any others)? The only realistic way I see, is to add up these counties & KCMO, and add their populations. But that only works if I have an confident list of counties that it spans :) (I would personally not use the 2010 census data btw, but rather the most recent census estimate) effeietsanders 03:55, 10 May 2020 (UTC)

effeietsanders, when it refers to Kansas City, they mean KCMO proper not and not the metropolitan area. KCMO spans the four counties that are listed and they report in those four also (see the latest report). Joplin also spans multiple counties (two) which you can see on their GIS map. I wouldn't use the JHU data for this as that data does not come directly from the health departments and is sometimes an estimate. The data that is used in the table/template is from the MO-DHSS page under the headings labeled "Cases by County or Jurisdiction" and "Deaths by County" respectively. As for the population data, I used official numbers from the last Census as it is the most accurate as estimates are just guesses where as the 2010 Census data was a confirmed count per se. — Mr Xaero ^☎️ 11:45, 10 May 2020 (UTC)

@Mr Xaero: From what I can see, the most reliable uniform source is JHU. One problem with taking everything from individual sources, is that you'll get messy definitions. JHU did a lot of the cleaning for us. For the purpose of creating uniform maps, I don't think I want us to duplicate that work. (In a way, going to all the individual government sources gets very close to primary sources) There's enough other messy edge cases to worry about without adding this into the mix. As far as I understand the census bureau numbers, although I wouldn't claim to be an expert, is that their annual estimates are much more likely to be accurate than the last full census. They make some extrapolations and assumptions, but I wouldn't call it a 'guess'. Even the 'full census' works with a number of assumptions and corrections. The added error you'll introduce by using 10-year old population estimates rather than 2-year old estimates is probably minimal though, so it doesn't matter too much. Note that in the census table, Joplin is not considered an entity (neither the JHU data nor the census mentions it separate). My best guess is that in the JHU data, most of its cases are assigned to Jasper County. Anyway, thanks for the insights, I'll see if I can include Missouri at some point in the coming days too. effeietsanders 20:35, 10 May 2020 (UTC)

effeietsanders, I am not quite understanding the logic of using JHU as the data source instead of the Missouri Department of Health and Senior Services (MO-DHSS). JHU obtains their data from MO-DHSS just like the CDC does. If you are questioning the data from MO-DHSS as being correct, then there are other issues as none of the data from any other source is correct as all others would be guessing the numbers and not providing the proper counts. JHU reports that they get their data from the following as per this quote taken from their website:

These include U.S. county and state health departments, multiple national government health departments, as well as data aggregating websites including 1point3acres, Worldometers.info, BNO,and the COVID Tracking Project (testing and hospitalizations), which rely on a combination of reporting from local health departments and local media reports.

I know that MO-DHSS might be behind a day or so as they have to gather from the local health departments, but the data on Wikipedia does not need to be realtime. — Mr Xaero ^☎️ 23:53, 10 May 2020 (UTC)

Just to be clear: I'm not saying you shouldn't use data directly obtained from MO-DHSS at all - there may be good uses for it. I'm saying that for the purpose of creating consistent maps across articles (which is what I'm working on), a consistent dataset is important. The JHU is exactly that. They aggregate their data indeed, and normalize it. I trust their experience to interpret data to mean the same more than my own. This is the whole reason why we try to use reliable secondary sources to begin with. The delay could be a factor, but doesn't really play a significant role for me. Much more important to me is that New York and Missouri maps are using roughly the same definitions. effeietsanders 00:34, 11 May 2020 (UTC)

Thanks for resolving. Alert Ted (talk) 15:19, 10 May 2020 (UTC)

I pushed an update of Missouri case counts. The population data is actually from Missouri's case count download. This matches the US Census Bureau's July 2019 population estimate for most counties. The separately broken out cities have their populations subtracted from the relevant counties. EphemeralErrata (talk) 09:45, 14 February 2022 (UTC)

References