(Un)designing Learning Agents in Messy Domestic Contexts: Lessons from a Grocery Prediction Algorithm Case - Grocery Receipt Database

This dataset contains transcribed grocery receipts from five households in Belgium and in the Netherlands for a period of eight weeks between November 2021 and March 2022 (split into two periods of four consecutive weeks). This dataset was used to design an experimental grocery prediction algorithm to predict shopping lists for 'week nine', for the purpose of researching its social and environmental implications for grocery shopping practices. This data collection included receipts from supermarkets, bakeries, butchers, organic shops, and any other sources through which the participating households bought their groceries. The dataset includes, for each purchased item: item name as mentioned on the receipt, quantity, price (per unit and total), store name (e.g., Spar, Carrefour), date and time of purchase, and whether there was a promotion on the item