Let data speak - but not for itself!
Data never just speaks for itself. If data speaks at all, it does with a strange voice, translating certain aspects of reality into a particularly standardised form – just like a piano translates certain activities into sounds. It is this formally shaped strangeness that makes up the particular value of data (and music). At the same time, that is what makes data problematic: These translations are only helpful if they lose information and reshape it in such a way that they are adding new information.