“Fabian Neuhaus author of Urban Tick at the Centre for Advanced Spatial Analysis, University College London has kindly agreed to write a guest post on the London Twitter network, the zoomable version embedded below is particularly notable:
Following up from the New City Landscape maps, they mapped tweet densities in urban areas around the world, we have now started to look into the socia network aspects of this data set. As a complementary graph to the map the network illustrates how the twitter users are connected through their activities and usage of the platform.
Graph by urbanTick / The London NCL Social Network graph of twitter users. The dataset is defined as geolocated tweets collected over the period of one week in the urban area of London set to a 30 km radius. Click on the image for a larger version on flickr or see the interactive zoomable version HERE.
The network is built from nodes and edges, were the nodes are the twitter users active during the time period of message collection back in May 2010. The edges visualise the connections between these users. From the messages sent connections are established based on activity and interaction. In reality these are the @ messages that are directed at one or more particular user. The second indicator of a connection are the RT messages, the message that have been retweeted by followers of the creator of the initial message.
Using these two methods the network graph is established as a directed network, meaning that the connection between the nodes has a direction since a message originates from a sender being delivered to a receiver.
The resulting network is built from a total of 17618 nodes and 26445 edges. In the case of this London twitter network not everyone is connected to everyone and about 5400 subnetworks were identified. Furthermore via the colouring the modularity of the network is visualised. Each subgroups has a unique colour shading indicating groups with tighter connections.”
text source: Digitalurban