Using approaches from artistic research, "Latent Spaces" develops aesthetics and performative practices to articulate the ambiguity of data as something productive. By understanding the meaning of data as open-ended, and by exploring how technical questions are inevitably also social and cultural questions that always include values and interests, it is possible to begin thinking about ways to negotiate this complexity.
Data is inherently ambiguous, that is, full of meaning to be decided upon, with an array of valid possibilities co-existing within ‘latent spaces.’ This ambiguity is produced and processed in every segment of the data pipeline, from the moment of data collection (or better, data creation), to the output of data-analytical processes, and everything in between. Drawing on the strength of artistic practices, we take this ambiguity as something productive because by understanding meaning in data as open, by exploring how technical issues are unavoidably also social and cultural ones, which always also includes values and interests, we can begin to think about ways of negotiating this complexity.
The research project unfolds through four artistic "field studies" and two transversal approaches. The first two field studies probe what is usually called “raw data,” here location data as it is emitted from beacon signals and from GPS systems, the latter two probe black boxes of data processing in social media. What all four have in common is that they use performative means to engage with and articulate the particular type of ambiguity of data in their respective field settings. The transversal approaches are data science and media theory. Both collaborate closely with the artistic field studies providing their specific tools but they also learn from the artists' ability to articulate ambiguity for their respective fields.