Raw Data Studies

Trying to remake a divorce percentage chart led me learn about the new Census Bureau table query interface to its Public Use Microdata.

Looking at the raw data behind two recent studies of early suicide counts during the Covid-19 pandemic gives me a chance to try out a “moving box plot” since the data is a bit sparse for regular confidence intervals.

The Southern Policy Law Center has released its “The Year in Hate 2020” report (PDF) cataloguing hate group activity throughout the year. I review some of the graphs. Perhaps a closer look will raise our awareness of the content.

Welcome to my new blog of my visual explorations of data! Some say raw data doesn’t exist, so you can think of “raw” as applying to the studies rather than the data if that helps.

Can beauty exist without meaning? I examine a time-based stacked area chart where the time component is of dubious value. Plus a rare pie chart appearance and a quick scatterplot remake.

A suspicious regression chart from an economics working paper leads to an exploration that finds an error in the original and some interesting alternate views.

It started with three questions about a chart. Diving into candle review ratings turns into a case for the “don’t smooth a smoother” rule.

A recent paper used DNA matching to support using a log function for dog-age to human-age conversions, but I’m not so sure the data supports it.

Remaking a puzzling bar-mekko chart in several forms. When width and height both encode data values, so must area.

A deeper dive into 93,000 insurance customer rate changes shared an analysis by The Markup. Making sense of suspicious residuals.