You are Biased and So is Your Data
Working with data is not just a science, but, to some extent, an art form – the art of storytelling. Data can tell stories to those who are eager to listen and they, in turn, tell their own version of the story. The story they tell can therefore obscure the facts due to the intended/needed outcome, limited understanding, preconceived ideas, biases, etc. But it is in their responsibility as storytellers to accept the data as it is and tell the whole story (no targeted information gaps). In order to do that we need first to become aware of some traps that can seriously sabotage our efforts to work correctly with data. Starting from the vocabulary we use around the data, the questions we ask ourselves, the spectrum through which we aggregate the data, how big the explored data space is, how we present the data, how we use the data in making decisions (single-loop learning or double-loop learning). As part of the presentation, I want to draw attention to some of these areas, with concrete examples from working at adidas Runtastic on the mobile app adidas Running. The examples would convey the areas in question but also suggestions on how to be self-aware of such things and how to react accordingly to them.