Fig 1.
Illustration of the modularity index.
In this fictitious configuration where 16 vertices form a network (panel A), we try to highlight some subgroups that are naturally communicate with one another. This scenario would naturally lead us to form four communities (panel B). The achieved modularity index is maximum (Q = 0.6071). Conversely, other configurations with 2 or 8 communities (panel C and D) reduce the Q index (0.4286 and 0.3578).
Table 1.
Comparison of the Edge Betweenness and Fast Greedy methods for analysing dialysis network activities in France.
Fig 2.
Results of the Fast Greedy clustering in the Aquitaine region.
The basemap was built from OpenStreetMap open data.: https://www.openstreetmap.org/copyright.
Table 2.
Description of the networks (Fast Greedy method).