Bubble artifacts
These tweets and other artifacts from my sociology/statistics bubble caught my eyes recently. I’ll try to publish similar curations of my bubble regularly from now on.
Fooled by visualization
Kieran Healy and John Mullahy show data is interpreted differently depending on how one chooses the scales of their axes:
They certainly would. See e.g. this pair (after an example by Bill Cleveland) where the perception that two lines are converging is driven entirely by the aspect ratio. https://t.co/e7huA05rGX https://t.co/gW2W38NLV5 pic.twitter.com/2CnRaQasss
— Kieran Healy (@kjhealy) April 19, 2020
Prediction is hard
And of course it’s not just difficult to not get fooled from choices of vizualisation, but it’s also not easy at all to predict s-shaped curves, as Constanze Crozier shows.
I spent a humiliating amount of time learning how to make animated graphs, just to illustrate a fairly obvious point.
— Constance Crozier (@clcrozier) April 17, 2020
“Forecasting s-curves is hard”
My views on why carefully following daily figures is unlikely to provide insight.https://t.co/yrE71bUXVT pic.twitter.com/BqQrxlhCmi
Merkel explains R values
Also related: Merkel explains the R-value and its consequence for society. It’s rare to see politicians explain network analytical measures from epidemiology so fluently. With English subtitles:
This is how Angela Merkel explained the effect of a higher #covid19 infection rate on the country's health system.
— Benjamin Alvarez (@BenjAlvarez1) April 15, 2020
This part of today's press conf was great, so I just added English subtitels for all non-German speakers. #flattenthecurve pic.twitter.com/VzBLdh16kR
Active learning for systematic reviews
Rens van de Schoot’s and Daniel Oberski’s team at the Methodology & Statistics Department of Utrecht University develop ASReview a wonderful tool which uses active learning to make systematic reviews easier. They even created a Covid-19 plugin. Cool stuff!
