From a Q&A with a VISA fraud prevention agent on reddit:
Some years ago, someone wrote a paper claiming he could get the age, gender and race only from the credit card purchase history. It worked very well. Today, with your full purchase information, we can even “guess” your income range, number of dependants and even weight. We have a statistical profile of every customer. We can even calculate the odds you eat at McDonald’s today, considering you ate there once every X days. 98% of the time, this model is very accurate.
One drawback is that it requires a lot of information. That is why it takes a few years and then, we are fully able to track you. In many cases, we compare the profile calculated from your purchase history to who you really are (and you thought they asked your income for credit validation) to further improve our models, and track fraud, most of all. It’s so sophisticated that if you order products a person in your group never ordered, your card will get automatically locked.
As I’ve mentioned previously, “I’ve always loved reading and learning aboutÂ data mining and its applications“â€”this is no exception.
From this Q&A I also discovered the Luhn algorithm.