Information Technology Reference
In-Depth Information
Table 11.1
Indices of
Density 6 %
Diameter 4
Average path length 2
Source: NBIC-Euro, Comin ( 2009 )
connectivity
the average path length of the graph, which is the average number of intermediate
connections between all pairs of vertices.
These three indices (Table 11.1 ) show that the connectivity of the graph is high:
although 22 % of all possible links are present, each city can reach all other cities in
an average of two and a maximum of four steps. Thus, the scientific-collaboration
network for converging technologies within European cities is strongly connected.
These results underline the interdependence of European cities for R&D in con-
verging technologies and provide promising information about the circulation of
scientific knowledge between cities.
Nevertheless, knowledge flows are not evenly distributed among European cities.
Rather, their patterns of accumulation are strongly affected by the pre-existing
population-size structure of the European system of cities ( Comin , 2009 ).
This phenomenon can be illustrated using the indices of centrality developed by
L. C. Freeman ( 1977 , 1979 ) to describe the relative importance of nodes within a
network. Two examples can be shown. First, the non-valuated degree centrality of
a city is the number of cities connected to it. This index measures the relational
activity of a city. Second, the betweenness centrality of a city measures its potential
intermediary role within a network: the more often a city occurs on the shortest paths
between other cities within the network, the higher is its betweenness centrality.
These two indices clearly distinguish the largest European cities from other cities.
Indeed, cities of the European megalopolis have large degree centralities (Fig. 11.2 ).
Tab le 11.2 shows that the 20 cities with the largest degree centralities are also the
largest European cities, except for two university cities: Utrecht and Goteborg. The
largest European cities, particularly capital cities, also have the largest betweenness
centralities (Fig. 11.3 ).
The structure of the scientific-collaboration network dedicated to converging
technologies reveals that European cities exchange knowledge through a classical
hierarchical-diffusion pattern ( Hägerstrand , 1952 ).
Moreover, our maps show that certain cities, such as Utrecht and Goteborg,
are more central than expected based on their population size. While large cities
concentrate the infrastructures that traditionally facilitate material and non-material
flows, specialized cities also have important innovation and training capacities. Spe-
cialized cities involved in European-funded converging-technologies R&D vary in
population size but are well connected to the network. Therefore, their hierarchical
positions in the network are much higher than their rankings among European cities
in terms of population size . For example, the empirical data analyzed here show
that Utrecht, Goteborg and Cambridge are ranked 11th, 19th and 24th, respectively,
in terms of degree and 84th, 92nd and 351st, respectively, in terms of population.
 
Search WWH ::




Custom Search