Charts and Graphs and Tables, Oh My

Posted by

OK, I might not have all three in this post. We’ll see. Just going to do my weekly round up of where I think things stand with the ‘rona. As with all data sources, take this with the appropriate grains of salt. As a reminder, I’m looking at three things: current cases compared to peak, current cases relative to population, and weekly change in cases.

States are given points on scales I developed and have held constant for a bit. One weakness is I haven’t factored in number of tests because you find more cases with more tests. The other caveat is I rely on state reported data and states are notorious for dumping cases on certain days (some of which are old) which skews the numbers.

As always, we start with the overall picture for the week that ended yesterday. The number of newly reported positives was 310,804 which is the highest number since the week of August 9th. While some of this reflects the last wave of states, some of it is driven by the large increase in tests. For last week, there were 6,280,580 tests reported which is the highest number of tests ever for a week. Even back at the peak of the sun belt wave, we didn’t hit this figure. Assume it is related to schools and colleges but don’t know. So the good news is that if you factor in the number of tests, the percentage of positive results was 4.9% which is down slightly from last week.

The lowest value I’ve seen is 4.4% back during the week of June 7th. The reality is we are never going to get to zero if we keep testing at this volume. There is a floor based on the false positive rate of the test which I can show later. Just assuming a 1% false positive rate, we’d estimate that 63,4000 (or 20%) of the “positives” recorded last week were really false positives. But as long as we are in a testing frenzy, we will keep finding cases.

Just to hammer that point a bit more, I did download data on tests (current to a week ago) and then looked at monthly results for a few places. Shown below is a table for New York. Viruses don’t really have this long tail. If you’ve been testing for almost four months and consistently finding 1% positive results, then you are really just finding false positives. You see similar patterns in places like Massachusetts, Connecticut and New Jersey. But we can’t let go of this obsessive need to test like crazy so we will continue to “find” new cases.


Last note on the overall stats, the number of newly reported deaths was 5,256 which is down from last week. We need a whole bunch of caveats around deaths. I use the term “newly reported” for a reason. It does not mean that 5,256 people died of the ‘rona last week. These numbers reflect death certificates which are notoriously slow to come in so you could even have deaths from back in July just showing up now. The CDC actually has a table tracking back to date of death which gives a more realistic picture.

For the week, we had 22 states that had higher scores than last week indicating some worsening of the numbers. There were 15 states that had lower scores. We tilt slightly towards things getting a little worse.

We start the first group with Vermont, Connecticut, New York and New Jersey with 5 points. Followed by New Hampshire, Arizona and Massachusetts with 7. Then we get Rhode Island and DC with 8 points and we close the group with California and Louisiana with 9 points.

Most of these are the northeast states where the virus really doesn’t seem to be an issue anymore. Arizona saw a small increase in cases but the weekly change is less than 2% so things are flat there. California and Louisiana are new (I think). Again, this is relative to population and not based on raw cases because in terms of raw cases, California is only behind Texas. But there are lots and lots and lots of people in California and new cases are growing at a small rate. It will just take a lot of time to get this to drop but clearly the pattern is in the right direction.

The next group starts with Maryland, Florida and Washington with 10 points. Behind them are Georgia, Delaware, Main, Nevada and New Mexico with 11 points. And we close the group with Pennsylvania and Hawaii with 12 points. Nevada and New Mexico are the losers in this group which Georgia made a big move in the right direction. Nevada and New Mexico saw increases in cases but the rate of change is less than 5% so it isn’t like they are back to rapid acceleration.

As for Georgia, here’s the graph. Down about half from the peak and you start to see some lower lows in the noise. In terms of cases per population, Georgia now ranks 6th in the country so wouldn’t expect things to start up again.

Now we come to the middle of the pack. Start with Ohio and Mississippi with 13 points. Followed by Virginia, Michigan and Oregon with 15 and then Tennessee with 16. We close the group with Indiana, Illinois, South Carolina and Alabama with 17 points. Mostly the big Midwestern states and the parts of the south where things haven’t dropped as much. Oregon is the big loser in the group based on an increase in cases. Still low relative to the population. Washington had some big problems early but it kind of seems like the virus just skipped the Pacific Northwest.

Virginia showed improvement in the numbers but I think that’s more an artifact of looking at one particular day. The case count is still high relative to peak but it has been flat since late July. This is another inconsistency of the virus because you get places with a long plateau like this. Currently, in terms of cases per people, Virginia ranks 33rd which suggests we could see this flat line extend for a bit. (Based on my assumption that you have to hit a certain amount of cases in order for this thing to burn out)

We start the next to last block with Alaska, Colorado and Texas with 18 points followed by West Virginia and North Carolina with 19. Then we get Minnesota and Idaho with 20, Kentucky with 21 and the group closes with Nebraska and Kansas with 22 points. In this group, Colorado, West Virginia and Minnesota currently sit at peak. For the group as a whole, the rate of increase ranges is roughly from 7% to 9%. I should have put this in some context but this is still not close to exponential growth. Texas, Colorado and North Carolina are the big losers in the group with case counts rising.

Here’s the picture for Texas. Relative to the other states in this group, it is still down from it’s peak. A lot of this is an artifact of the one day with almost 18000 tests. I couldn’t find anything to explain that other than a dump of some past data. That one value will be out of the average next week and, unless we get another spike, Texas should move back down the list.

And then we get to the bottom of the list. We start with Iowa with 23 points, followed by Missouri and Wyoming with 24. Then we get Arkansas with 25 points followed by Oklahoma and Utah with 26. We close with Wisconsin and Montana with 27 points and North and South Dakota with 28.

Montana saw cases increase by 17% last week which is the highest rate of change in the country. Here is what that looks like. In terms of cases per people, Montana ranks 43rd in the country so this could really be the beginning of big move up. We’ll know more next week.

As a bonus, here is the latest graph for North Dakota. One of the things I do is try to predict a week in advance. When my prediction is consistently lower than the actual number, it indicates a place where things are still going up. As things start to flatten, the predictions becomes more accurate. That is the case for North Dakota. It is not yet reflected in the 10 day average but it looks like the acceleration phase is over and we should see this plateau. Since North Dakota currently ranks 11th in terms of cases per population, I would expect to see this start to drop fairly soon.

And that’s my opinion of where things stand as of yesterday. To me, the two states where things are still not yet in control are Montana and Wyoming.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.