Introduction - Methods for Multilevel Analysis and Visualisation of Geographical Networks - page 6

Fig. 1.2 The multilevel organisation of cities

social phenomena through the powerful and extremely relevant notion of level.
Social relations naturally contribute to multilevel network dynamics. Although the
view that the population dynamic simultaneously acts at the individual and group
levels is plausible, formulating and validating a multilevel model is a painstaking
endeavour. From the computer science perspective, the “level” concept provides
strategies for downsizing data and potential methods for handling extremely large
datasets. Modelling networks and flows as weighted graphs involves the use of
graph theory, which transforms a multilevel network into nested subgraphs. Actors
or entities interacting at a lower level establish higher-level entities that interact with
each other, and flow exchanges at lower levels establish higher-level flows.
Incorporating cities into multilevel networks appears to have occurred not only
in the field of transportation but also in the area of social and organisational
networks ( Dicken , 2011 ; Grabher , 2006 ). Although agent networks transcend the