The pedestrian trace is not an innocent line deposited upon an otherwise settled map. Larroya and colleagues release roughly 3,000 geolocated records from 19 trajectories produced by 72 participants working in 19 groups across three communities, alongside processed data and explicit accounts of protocols. Participants did not merely travel from origin to destination: they selected stopping points and carried out collective, sometimes festive, social actions. The dataset consequently joins movement to the production and interpretation of urban situations. Its distinctive contribution is infrastructural, making raw and filtered observations available with enough methodological context to support further analytical work. Between citizen science, computational urban research and spatial design, it suggests that data publication is itself a continuation of the research apparatus: filtering, precision and documentation determine what can later be compared. Yet an exploratory event-based trajectory cannot stand automatically for habitual behaviour or longitudinal change. The operational lesson is to keep route geometry, participant action, field context and processing history jointly accessible, so that a clean spatial visualisation does not erase the conditions of its production.