Françoise LUCCHINI, Univsersité de Rouen, France
Bernard ELISSALDE, Univsersité de Rouen, France
Julien BAUDRY, Université du Havre, France
To apprehend the variety of urban mobility patterns within the urban space, geo-localised digital data derived from different urban detection systems and/or connected devices is now used. The traces left by individuals by way of these devices have become an essential data to understand the functioning of the urban space as a whole (mobility flows, concentrations) or to capture specific behaviours on the occasion of special events. From the information provided by geo-localised APIs data from micro-blogging Twitter, we chose to explore the patterns of mobility behaviours of users of this social medium in Paris intra muros through the years 2015 and 2016. By way of the precision of the information, and its variability over time in different parts of the Paris urban space, mobility phenomena can be brought to light. We attempt here to analyse apparent movements using index calculations belonging to graph theory, and we set out to identify the main polarisations and trajectories using the indicator provided by the degrees of the vertices. Positioning this work in the scientific field of complex networks, and following the seminal works of A.L.Barabasi, we will endeavour to apprehend the relative share of the variant and the invariant in weekly graphs derived from the clustering algorithm ST-DBSCAN used to process Twitter data. The results obtained yield a picture of the patterns of movement of Twitter users in Paris that is at once probable and uncertain. The mobility graphs obtained from the recordings of the Twitter social network thus present patterns of complexity on several levels. The most frequented vertices are markedly polarised and almost invariant in the localisation. The vertices with middling and low degree values are partly random, they do not necessarily follow a decreasing hierarchy for the degree values. Connected devices tend to put the emphasis on what is happening (personal opportunity) over what is permanent and static in the city.
Mots clés : urban area|mobility|graphs|complex systems
A103327FL