"Data double" is a term increasingly used to describe (a) the now-widespread practice of constructing individuals' "profiles" on the basis of aggregated data collected from a wide variety of sources, and (b) the problematic relationship between such profiles and the "real" person. Data doubles are widely used both in commercial contexts - as when Facebook or Google constructs a profile of us and commercializes it for developing advertisement strategies - and in security and intelligence contexts, where "profiling" is increasingly based on gathering and aggregating digital data through hi-tech surveillance tools.
It is important to realize that data doubles - whether or not they correspond to the "real" person - are "real" in themselves. A data double can be used to cast suspicion on "real" people, to trigger discrimination and exclusion. When your data double shows a socially unwelcome profile, you are in "real" trouble.