Thread of quick hit thoughts about the Zvi blog post from me, a data scientist. I read his post from today 10/29 and the one prior to it from 10/22.
I thought the posts were interesting, but ultimately they are presentations of data followed by some subjective opinions. To someone that isn& #39;t data literate, it& #39;s easy to pass those opinions off as hard facts / obvious conclusions from the data. They aren& #39;t.
Data analysis is best meant to augment subject matter expertise, not replace it. I don& #39;t think Zvi was trying to talk over experts or cast himself as one, but I wouldn& #39;t buy too much into anything he says that doesn& #39;t get an explicit stamp of approval from those experts.
Case in point, one of my data jobs was in anti-money laundering. I had access to billions of transaction records, and could have maybe found some potential instances of laundering, but it wasn& #39;t until I got certified in AML (CAMS) that I realized how little I knew about it.
I think casting pandemic-cautious people as the "doom patrol" is extremely harmful. If you aren& #39;t exactly data literate and read a post like that, an easy conclusion you could potentially draw is "hey, a smart data guy says it& #39;s possible to take the virus TOO seriously".
That could convince someone to relax their risk tolerance for virus exposure, which they might not have a good grip on how to do. Enough instances of that WILL lead an increase in infection (which we all know the consequences of). To me that is unbelievably irresponsible.
I am just as fatigued with masks and distancing and etc. as you are, but absolutely do not let a blog post with a bunch of charts and graphs somehow convince you to let your guard down.
I also saw some replies dismissing Zvi and Zvi-defender arguments b/c they came from cis white men, or b/c "data science" is tech bro nonsense (it isn& #39;t), or b/c they were friends with <horrible Magic person>. Cut this shit out. These aren& #39;t productive arguments and you know it.
Don& #39;t overcalibrate someone& #39;s arguments because of their identity. Smart people can be misguided & lack knowledge, but they are still allowed to be smart.
To summarize, I think the posts could have been mostly good if they stuck to just facts, but ended up being bad for a small variety of reasons (the subjective nature of the posts + the potential influence they carry).
You can follow @AlexMajlaton.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: