Multidisciplinary Research on Geographical Information in Europe and Beyond Proceedings of the AGILE'2012 International Conference on Geographic Information Science, Avignon, April, 24-27, 2012 ISBN: 978-90-816960-0-5 Editors: Jérôme Gensel, Didier Josselin and Danny Vandenbroucke
Geographic Analysis of Social Network Data
Chris Brunsdon Michael Batty, Alexis Comber University of Liverpool Andrew Hudson-Smith, University of Leicester Fabian Neuhaus & Leicester, UK, LE17 7RH Liverpool, UK, L69 3BX firstname.lastname@example.org Steven Gray, email@example.com University College London, UK, W1T 4TJ Abstract
This research analyses social network data to identify communities or sub-graph regions. These sub-graph areas are indentified based on the arrangement of edges between vertices. The geographies of the communities are analysed, compared and visualised using kernel density estimations. A research agenda is suggested. Keywords: Graph Theory, Network Communities, Sub-Graph Geography, Twitter Data, London
firstname.lastname@example.org email@example.com firstname.lastname@example.org email@example.com
This paper introduces methods for the spatial analysis of social network data. Social network increasingly has a geographical component and it is possible spatially analyse sub-graph geographies. The paper describes methods for the identification of sub-graph regions that represent communities and mapping their spatial extent. It draws from research in statistical physics for partitioning networks in order to identify ‘communities’ or areas of the graph that are homogenous in some respect and from classic spatial analysis. In so doing it addresses recognised concerns over the reliability of the communities that are identified using these methods and the difficulty in understanding what they mean  .
Social Network case studies
Real networks tend to be irregular and highly heterogeneous, with specific parts of the network or graph (the terms are used interchangeably here) having high concentrations of interconnected vertices. The aim of community detection is to identify areas of the network that have high concentrations of edges that connect groups of vertices and that have low concentrations of edges between these groups. Such areas can be considered as ‘communities’  Methods have been developed for partitioning networks in order to identify communities – areas within the graph (sub-graph areas) where the nature of interactions between vertices indicates some local clustering of interactions, under the assumption that subgraph areas with high internal interactions are homogenous to some degree, depending on the nature of the network (social, publishing, cell phone etc). The interested reader is directed to number of reviews of the methods arising from statistical physics   . Recent work in the geography literature indicates that some community detection methods are more suitable for geographical applications than others because of the inviolable nature of topological network properties . The case studies presented here identify methods for partitioning social network data into sub-graph areas and for examining their geographies.
Data was collected for an area of 30 km radius with its centre in Parliament Square in London. For each record (tweet), a number of items from the metadata of the message are returned including: • The username of the sender. • The content of the tweet. • The time the tweet was sent. • A geographical location the tweet was sent from. In the case studies specific tags in communications between social network data users (‘@user’ in Twitter data in this case) are used to identify and illustrate the connectedness of different concepts. The network is defined by the interactions (edges) between users (vertices). A subset of the data was analysed. It contained 87,555 records. Of these 52,397 contained tweets at (‘@’) a specific user, 52,280...
References:  Porter, M.A., Onnela, J.-P. and Mucha, P.J., (2009). Communities in Networks. Notices of the AMS, 56(9): 10821166.  Newman, M.E.J., (2008). The physics of networks, Physics Today, 61(11): 33-38.  Girvan, M. and Newman, M.E.J., (2002). Community structure in social and biological networks. Proceedings of the National Academy of Sciences, 99: 7821-7826.  Fortunato, S., (2010). Community detection in graphs. Physics Reports, 486(3-5): 75-174.  Leicht, E.A. and Newman, M.E.J., (2008), Community structure in directed networks, Physical Review Letters, 100: 118703.  Comber A., Brunsdon, C. and Farmer, C. (in press). Community detection in spatial networks: inferring land use from a planar graph of land cover objects. Paper accepted for publication in International Journal of Applied Earth Observation and Geoinformation (January 2012)  Newman, M.E.J and Girvan, M., (2004). Finding and evaluating community structure in networks. Physical Review E, 69: 026113.  Pons, P. and Latapy, M., (2005). Computing communities in large networks using random walks. http://arxiv.org/abs/physics/0512106v1.  Venables, W. N. and Ripley, B. D. (2002). Modern Applied Statistics with S. Springer, New York, Fourth Edition.  Schaefer, D.R., (2012). Youth co-offending networks: An investigation of social and spatial effects. Social Networks, 34(1): 141-149  Takhteyev, Y., Gruzd, A. and Wellman, B. (2012). Geography of Twitter networks. Social Networks, 34(1): 7381.
This research introduces statistical methods for analysing communities in social network data and their geographic extent. These that provide greater insight into social network structure, content, associated concepts and their geographical aspects. A research agenda is suggested as a result of these initial analyses.
Multidisciplinary Research on Geographical Information in Europe and Beyond Proceedings of the AGILE '2012 International Conference on Geographic Information Science, Avignon, April, 24-27, 2012 ISBN: 978-90-816960-0-5 Editors: Jérôme Gensel, Didier Josselin and Danny Vandenbroucke
Please join StudyMode to read the full document