Major revisions to format, data types, and more.

With The Covid Tracking project decommissioned (don’t worry, you can still access all of their data), there have been a lot of questions about what I plan to do next with my own data resources and interfaces.

The CDC now provides more data about what’s going on in Florida than Florida’s own Department of Health (DOH), and that data often shows a much more grim picture than the state would lead people to believe.

For example, hospitalization data provided by the Department of Health and Human (DHH) Services showed 3,725 beds in use for COVID-19 patients, while the Florida Agency for Healthcare Administration (AHCA) showed only 2,813 (data as of 3/14 at 10:46 AM EST).

And DOH of course does not even include AHCA data on the dashboard I built for them more than a year ago, or data from the Florida Department of Emergency Management (DEM), or the Florida Department of Corrections (DOC), or any other state agency reporting COVID-19 data to DOH – the result of a juvenile “turf war” between DOH and every other state agency as to whose data would be highlighted on the website.

But hospitalizations aren’t the only place where the numbers disagree.

Testing data is another issue. According to the CDC, Florida testing volume has been on a gradual decline since the start of 2021, down 17% – a trend consistent with what we’ve observed in a number of states.

But you wouldn’t know that if you just looked at Florida’s website, which uses the way it displays data to make it appear that testing figures have been relatively unchanged. This goes back to the “how to lie with maps” discussions and selectively displaying data in a way that creates a false illusion of progress.

So this leaves me asking the question: who’s data is better? The feds, who collect part of their data from DOH, but also do their own surveillance data to “check” states like Florida with reputations for fudging numbers? Or do we continue with the plan to only include official Florida data produced by all of its respective agencies?

I’m leaning feds, but would appreciate your feedback!

10 thoughts on “Major revisions to format, data types, and more.

  1. Thank you so much for keeping this up! Florida has done an extremely poor job – both, of keeping the public aware and also keeping it’s citizens protected.

  2. Please, continue to use the Federal numbers. Online information should capture the most accurate data possible. Deeply appreciate all your efforts!

  3. Thank you for continuing due diligence. The vaccination roll out here in Florida is also a disorganized nightmare.

  4. I think incorporating cdc info into your website is a great idea, especially if you are looking to highlight the differences.

    Though there are important distinctions to make with the data: cdc uses only pcr tests in their testing numbers and positivity calculations. I haven’t run an analysis, but are we seeing a jump in Antigen testing as well as a dip in overall testing? That could be interesting to see as a visual.

    With the HHS data vs AHCA, primary diagnosis of covid vs. all people with positive covid tests – distinctions the state made to make it seem better than it actually is. That drives me bonkers.

    I also really like the age breakdowns of hospitalizations in the community profile report from cdc. I’ve been really paying attention to that since the vaccine rollout.

  5. Showing both side by side is enlightening. It’s the “unknown unknowns,” though, that have bothered me all along. I have always felt that the testing numbers, in particular, bare no relation to reality. How many people have been discouraged from getting tested at all? What about the absence of contact tracing and the cases that are missed there? Etc. Etc.

    I do appreciate the work you have continued to do, and your dashboard is a daily ritual for me. Thanks so much.

  6. Thank you so very much! You have been such a gift to Floridians who believe in honesty and science. Your efforts are truly appreciated!!

Leave a Reply to Patrick Kitts Cancel reply

%d bloggers like this: