Data mining involves four phases: collection, aggregation, query and prediction (de Zwart et al 2014, p. 715-719). As with filtering, this process most often occurs without users' knowledge, and regardless, it is "nearly impossible" to prevent (p. 716). The importance of user consent has already been addressed, as have the implications for the balance of power between social media platforms and users. Since social media implicitly exists to allow users to interact with each other, the power to remain out of users' hands is particularly problematic as it contradicts in many ways the supposed purpose of social media (Baym 2015, p. 1-2). . More specific to data aggregation and interrogation, power imbalance works against freedom of expression and instead limits an individual's ability to control their identity and how it is presented. Data aggregation, or the collection of data over time, has no way or context of retention for specific actions or interactions (de Zwart et al 2014, p. 716). The context collapse discussed above once again comes into play, as how an individual may present themselves in interactions with various other users can form seemingly contradictory information. The query of this data then becomes the