Pedestrian mobility features of a citizen science experiment for street-level urbanistic interventions

We present preliminary results of a citizen science experiment that aimed to revise the perspective of a well-defined public space of the "Barri Primer deMaig" neighborhood (0.03km2) in the city Granollers (Catalonia). The neighborhood shows a very particular and somewhat problematic urban structure in social terms. To better understand current patterns in public space use and to identify potential interventions to improve the quality of the neighborhood public space, we have designed a pedestrianmobility experiment.

More particularly, we havestudied 18 purposed-specific pedestrian journeys corresponding to groups of 3-5 people with different socio-demographic characteristics. The groups of participants were asked to complete a variety of missions in which they have to choose particular sites to complete a list of activities to be made in public space: dance, play football, have a snack, drink a toast, etc. We have used the Wikiloc application to collect 2, 981 GPS locations. A set of data visualisations on maps has allowed us to identifymovements flow areas or the location and duration of stops to accomplish the set of missions. Locations identified have a strong interest for urban planning in a street level. However, other analysis are also possible. Exponential decay and a characteristic time scales can be identified in stops duration. Themean squared displacement (MSD) has the ability tomeasure the type of diffusion of the groups, which in our case show a super-diffusive regime, almost ballistic (fully oriented, with no dispersion and with fairly constant velocity). For this reason, we therefore have opted to focus the analysis on instantaneous velocities which shows a randompath with reversing fluctuations towards a certain normal velocity level. The probability density function of all instantaneous velocities is observed to be consistent with a Log-Normal distribution and shows full coherence with previous citizen science experiments we performed in 2018. Furthermore, the observed reversing fluctuations around a given velocity value can be analysed in terms of the auto-correlation, which shows an exponential decay as well. Given thesemobility statistical patterns, the logarithmic Ornstein-Uhlenbeck process appears to be a good candidate formodeling the pedestrian movement in these particular journeys as we have found in other experiments.