obra’s Twitter Archive—№ 10,041

The canonical 1965 study of English letter frequency used a 20k word corpus. Norvig reran it on Google books data: norvig.com/mayzner.html