January 11, 2013, Webb 1100

We describe several recent results on large network analysis with a special emphasis on community detection and on the analysis of mobile phone datasets. In particular, we describe two simple and efficient methods - the "Louvain method" and the recent Partition-Merge method - for the detection of communities. Partition-Merge is a general and versatile method that allows to obtain a distributed version of any centralized algorithm for graph computation while keeping accuracy guarantees. The Louvain method - now used by LinkedIn for its visualization application InMaps - has sub-quadratic computational complexity and can be routinely used for analyzing networks with billions of nodes or links. We analyze communities obtained on a nationwide dataset of criminal records, as well as on a social network constructed from mobile phone communications in Belgium and in France on periods covering several months. In Belgium we observe a gravitational law of social interaction as well as spatially distributed social structures that have potential political implications. We also describe applications of mobile phone dataset analysis for a range of applications such as urban planning, traffic optimization, monitoring of development policy, crisis management, and control of epidemics. With these applications in mind we overview the ongoing "Data for Development" (D4D) challenge (http://www.d4d.orange.com/) organized jointly with Orange on the analysis of mobile phone datasets from an African country and for development purposes. About 200 research groups are currently participating in this challenge.

Vincent D. Blondel is professor and head of the department of mathematical engineering at the Université catholique de Louvain (UCL) in Belgium. He received an engineering degree, a degree in philosophy, and a PhD in Applied mathematics, all from the Université catholique de Louvain, and a MSc in pure mathematics from Imperial College (London, UK). He has also completed a master thesis at the Institut National Polytechnique de Grenoble (France). He was a visiting scientist at Oxford University in 1993. During the academic year 1993-1994, he was the Göran Gustafsson Fellow at the Royal Institute of Technology (Stockholm, Sweden). In 1993-1994 he was a research fellow a the French national research center in computer science (INRIA, Rocquencourt, Paris). From 1994 to 1999 he was an associate professor at the Institute of Mathematics of the Université de Liège in Belgium. Dr Blondel was a visitor with the Australian National University (1991), the University of California at Berkeley (1998), the Santa Fe Institute (2000) and Harvard University (2001). He has also been an invited professor at the Ecole Normale Supérieure in Lyon (1998) and at the University of Paris VII (1999 and 2000). In 2005-2006 he was an invited professor and
Fulbright scholar with the Department of Electrical Engineering and Computer Sciences of the Massachusetts Institute of Technology(Cambridge, USA).
 
Dr Blondel's major current research interests lie in several area of mathematical control theory and theoretical computer science. He is a former associate editor of the European Journal of Control (Springer) and an associate editor of Systems and Control Letters (Elsevier) and of the Journal on Mathematics of Control, Signals, and Systems (Springer). For his scientific contributions he has been awarded of a grant from the Trustees of the Mathematics Institute of Oxford University (1992), the prize Agathon De Potter of the Belgian Royal Academy of Science (1993), the prize Paul Dubois of the Montefiore Institute (1993), the triennal SIAM prize on control and systems theory (2001), the prize Adolphe Wetrems of the Belgian Royal Academy of Science (2006), and the Antonio Ruberti prize in systems and control of the IEEE (2006).