[lazzlazzlazz]

> details of what they like/would likely click on/spend money on

These details can only be extracted with very complex natural language processing techniques.

> where they are in the world

The vast majority of Twitter users have no structured geographic location.

> what age they are, what gender they are

Twitter doesn't even have fields to enter age and data. It's all inferred, which makes it difficult to extract accurately (if at all).

[otterley]

Twitter, like Facebook, has beacons on just about every website out there, thanks to website owners who are all too happy to make it easy for visitors to click "like" or "tweet" in the hopes of attracting attention. You don't even have to post a single Tweet for Twitter to know about what you're into.