A traditional data investigation cycle includes problem posing, planning and collecting data, analysing data, and making conclusions. This research studies the data investigation cycle for analysing image-based data. In three independent research projects, students at different educational levels and from different countries were provided photographic data of families and their environments around the world from the Dollar Street project. Data collected included classroom video-recordings (Australia), individual student interviews (Colombia), and pre-service mathematics teachers’ interviews (Turkey). Analysis focused on the sequence of actions that helped students when attempting to pose and answer questions based on the data set. Findings suggested a similar, iterative sequence of actions across all cohorts: context and data set familiarisation, variable identification/generation, problem posing and planning, data organisation and analysis, and drawing conclusions.