Publication Details

Download

Florian Bemmann, Daniel Buschek
Interaction Challenges for N-Of-One Experiments based on Mobile Sensing Data
Workshop on Grand Challenges in Personal Informatics and AI @ CHI 2022, May 11, 2022, Online (bib)

Mobile Sensing data collected by smartphones contains a rich body of information. Researchers and companies already leverage that for research and businesses, respectively. However, the end-user from whom the data originates is not enabled to do so yet. We show how smartphone applications could enable users to leverage their data themselves. In N-of-1 experiments users conduct research questions on themselves. By passively tracking behavioral and contextual data, enriching it with a self-reported variable of interest (e.g. mood, stress, cognitive load), and using appropriately presented statistical and/or Machine Learning methods they could reflect on and learn about the factors that might influence personal variables of interest. To make progress towards this vision of empowered end-users, we outline a Mobile Sensing app for N-of-1 experiments, point out relevant challenges, and propose ideas and future research directions to solve them. More broadly, we envision mobile sensing-based machine learning applications to spread in the area of personal informatics.