Association has become a central aspect of surveillance and a key practice of making information matter. It is critical to any kind of profiling that we experience on an everyday basis. To associate is to join, to make a connection ‘in an interest, object, employment or purpose’ (Harper, nd). One of the most widespread ways of analysing information is indeed to make a connection between different datasets. In her work on data derivatives Louise Amoore speaks of an ‘ontology of association’ (2011: 27). This means that associating data is not just a knowledge practice, but it describes a specific way in which data materialize and come to exist together. The most common approach of associating different datasets with each other follow a Boolean logic (Kitchin, 2016), named after the mathematician George Boole. We know them as if-then rules, that is: when if is true, then is executed. By means of if-then instructions disaggregated data are associated ‘to derive a lively and alert new form of data derivative – a flag, map or score that will go on to live and act in the world’ (Amoore, 2011: 27). The aim of associative practices is to connect different sets of information and to derive patterns from them (Kaufmann, Egbert, and Leese, 2019). What is more, such patterns, again, are likely to be associated with actions that matter to society. Whether a pattern is considered meaningful and actionable depends on many aspects, not least those involved in associating. Association is a classic analytic practice, which is also used to process analogue information. With the rise of digital information, however, association has shifted in terms of reach and quality.