Sunday, April 19, 2020

Projections

After some frustration with the IMHE model, my favorite statistician, bsking, went looking and found a UTexas site that shows the results of ten different models, illustrating the current uncertainty and dependence on assumptions. She works at Dana Farber Cancer Institute and has to fight through health statistics to earn her daily bread.
at

4 comments:

  1. bs king1:36 PM

    LOL, probably good you distilled my poorly spelled and poorly formatted email rant down to "some frustrations". These migraine meds do make it harder to put longform thoughts together.

    ReplyDelete
  2. RichardJohnson2:51 PM

    Note that they threw out the April 14th 6,185 deaths that NY state added for a correction. Makes sense to me.
    Daily New Deaths in the United States

    ReplyDelete
  3. james4:31 PM

    The projections do sort of cluster, don't they?

    ReplyDelete
  4. RichardJohnson10:24 PM

    I have been graphing various Corona/Winnie the Flu stuff for about a month. One graph is (total deaths today)/(total deaths yesterday). Throwing out the April 14 outlier, from March 23 (1.39) onward,drawing a line on the peaks (but not April 14), the decline is fairly consistent to April 19. (1.04).

    ReplyDelete