Basic analysis: interpretability
• The nodes in our network correspond to real entities. For each place in the
network, represented by its id, we have its title and geographic coordinates.
• Iterate through the lists of centrality nodes and use the meta data to print the
titles of the respective places.
### READ META DATA ### node_data = {} for line in open( ' ./output/cambridge_net_titles.txt'): splits = line.split(';') node_id = int(splits[0]) place_title = splits[1] lat = float(splits[2]) lon = float(splits[3]) node_data[node_id] = (place_title, lat, lon) print 'Top 10 places for betweenness centrality:' for node_id in top_bet_cen: print node_data[node_id][0] 35
Basic analysis: most central nodes
• The ranking for the different centrality metrics does not change much, although this
may well depend on the type of network under consideration.
Top 10 Cambridge Railway Station (CBG)
Grand Arcade
Cineworld Cambridge
Greens
King's College
Cambridge Market
Grafton Centre
Apple Store
Anglia Ruskin University
Addenbrooke's Hospital
Top 10 Cambridge Railway Station (CBG)
Grand Arcade
Cineworld Cambridge
Apple Store
Grafton Centre
Cambridge Market
Greens
King's College
Addenbrooke's Hospital
Parker's Piece
Top 10 Cambridge Railway Station (CBG)
Cineworld Cambridge
Grand Arcade
King's College
Apple Store
Cambridge Market
Greens
Addenbrooke's Hospital
Grafton Centre
Revolution Bar (Vodka Revolutions)
Betweenness centrality
Closeness centrality
Eigenvector centrality
36
