According to the data, the average Twitter user is female and in her late teens. She follows between 20 and 50 people and has about the same number of followers herself.
The new study from British web consultancy firm Box UK collected data during July and August from a random selection of 83,628 Twitter users. The criteria for separating "real" from "spam" accounts set a threshold of more than 5 tweets, 5 friends and 5 followers, and had to have changed their avatar photo.
The results show the data from both the "real" group and "all users," lending some insight into the typical patterns of spam accounts as well. Of all the methods described, perhaps the most variability might be found in the age determination, which was taken from phrase mentions that included age but not overt mentions of birthdays. The data still shows skews from important milestone birthdays and there's also undoubtedly a selection bias from older users being less likely to publicly announce their ages.
Still, with all collection methods described and grains of salt taken, the data paints an interesting picture of the average Twitter user as a girl in her late teens. Some interesting patterns regarding spam accounts could potentially help improve Twitter's own algorithms -- or perhaps third parties could take advantage as well -- for identifying and rooting out spammers.