Constitution of Databases: The (Un)Predictability of Food Waste

Apps that aim to reduce food waste by redistributing it are an increasingly popular way of countering excess food. In this paper, I explore the case of an app in the making that focuses on altering grocery shopping behaviour to reduce household food waste. I ask how the constitution of its database reconfigures practices of food waste reduction. ‘Ground truthing’, defined as the constitution of databases (Jaton, 2017), helps me to frame the development team’s valuation processes around their data gathering. After the literature review, I introduce my case, WasteMentor, a research-based app with the goal of conducting digital receipt-based sustainability and food waste monitoring and interventions. This app builds on grocery shopping data collected through customer loyalty programs of two major Swiss retailers as well as a ‘food waste collection study’. I have taken an in-depth ethnographical approach to studying the development of the app’s infrastructure (Star, 1999; Hine, 2015; Schneider & Eli, 2021). Analysing how the discarded food is turned into data and combined with digital receipts to eventually reach the ground truth, I argue that the narrow focus on mainstream retailers neglects how food (waste) is enacted in practice. I conclude with a critical reflection on the attempt to datafy and predict (potentially) decaying food, whereby I aim to contribute to critical infrastructure studies.