Graphical Error (or Deception)?

There has been a couple of interesting articles lately about how pandemic data is being collected or depicted. These articles are illustrative of how we should be thinking about the data or graphs we are seeing. Georgia had the most blatant error in handling data. Or was the graph a deliberate deception?

The following are graphs I tried to create for Georgia: the one on the left shows confirmed cases in declining order (not in order of date) and the one on the right has the addition of the right kind of graph in the upper right hand corner.

(Sorry, dates are hard to see. There is actually more dates extending out to the right but ran out of room. The orange bars are for 5/21 and 5/22.)

In my opinion, I find it hard to think the graph was a sorting error. It’s very hard to make the dates go out of order – I think most programs just automatically sort by dates. I had to apply some fancy tricks to make Power BI sort by confirmed cases (largest to small) rather than sort by date. It took me a while to generate a Georgia graph showing declining cases, with dates out of order. I’m inclined to think Georgia’s error really was just an out and out deception. So, when you look at a graph with steady declining (or increasing) numbers, check the dates.

The combination of tests for infection with tests for antibodies is easier to explain. I can imagine that the reporting systems, when developed by IT, would have only one column or input box for testing data, thus forcing those reporting testing results to the states to combine the infection tests with the antibody tests. It’s very conceivable that the IT programmer would not think of multiple types of tests when initially designing the collection program. We’re still in the early stages of this pandemic. Fortunately, the experts, such as the scientists and doctors were on to this problem and I hope the states will correct this issue. (I wish some of those science deniers would wake up!)

But you can see how easy it is to get derailed when trying to rely on data, information or graphs. It’s always important to scrutinize or to think of what could go wrong or what could be missing.