Database Reference
In-Depth Information
Figure 10.14 Visualization of the clusters determined from the mobility network. As a
reference to the existing administrative borders, the perimeter of each town is drawn with
a thicker line. The clusters are determined considering the communities derived from
Infomap: regions within the same cluster are themed with the same shade.
border of the city. This misclassification is not necessarily a symptom of a error,
but rather proves that the sector is attracted by the adjacent city.
Another relevant property of the clustering results is an empirical proof that
a single city cannot be considered an “island.” On the contrary, the mobility of
a city strictly depends on the mobility of the surrounding towns. In fact, each
cluster can be described as an enumeration of a series of cities. Moreover, the
cohesion property described above allows the definition of a partition of the
territory where each group, that is, each cluster, can be exploited to develop
combined mobility policies and planning. Finally, it is important to note that all
the clusters present geographically adjacent census. Although this constraint has
not been imposed to the community discovery algorithm, the strong cohesion
of regions is yielded by the high volume of internal mobility that enables the
regions to attract each other. Thus, as discussed earlier, the local short-ranged
movements dominate the long-range trips.
