Data gathering and tabulation are the historical and technological basis of what we do. Yet I believe the fundamentals have come to be neglected, in favour of calculations which convince no one and plots which entertain but don’t teach.
Here are two tabulations I’ve come across this year. There are no plots. The only calculation involved is cat.
As far as interpretation, the reader is expected to understand the meanings of the columns herself, and do some thinking of her own. No questions about sufficiency, width, or slope are posed.
And yet these tables communicate to the human beings who read them. Because there are no further derivations, the important questions about provenance and truth are closer to the foreground, although I don’t believe the data gathering is very problematic.
Number one: Mother Jones’ google doc of mass shootings in the United States.
(you can compare this to Nate Silver’s outfit’s attempt—which had plots)
Two is Irvine Sprague’s list of FDIC assistance to US banks & thrifts.
The polar opposite would be Diane Griffin Saphire’s Estimating Victimization Prevalence. (https://books.google.com/books?id=nC3UBwAAQBAJ&lpg=PR2&dq=estimating%20victim%20prevalence%20diane&pg=PR2#v=onepage&q&f=true)
A non-statistician might ask how it’s possible to bootstrap mathematically something that is going unreported in reality. After reading the book, I don’t think it is possible to do so — but one can certainly post charts and discuss competing models at length, apparently at enough length to complete a book.