Guy Melancon (Editor)
This leading-edge study focuses on the latest techniques in analysing and representing the complex, multi-layered data now available to geographers studying urban zones and their populations. The volume tracks the successful results of the SPANGEO Project, which was set up in 2005 to standardize, and share, the syncretic, multinational mapping techniques already developed by geographers and computer scientists. SPANGEO sought new and responsive ways of visualising urban geographical and social data that reflected the fine-grained detail of the inputs. It allowed for visual representation of the large and complex networks and flows which are such an integral feature of the dynamism of urban geography. SPANGEO developed through the ‘visual analytics loop’ in which geographers collaborated with computer scientists by feeding data into the design of visualisations that in turn spawned the urge to incorporate more varied data into the visualisation. This volume covers all the relevant aspects, from conceptual principles to the tools of network analysis and the actual results flowing from their deployment. Detailed case studies set out in this volume include spatial multi-level analyses of flows in airports and sea ports, as well as the fascinating scientific networks in European cities. The volume shows how the primary concern of geography—the interaction of society with physical space—has been revivified by the complexities of new cartographical and statistical methodologies, which allow for highly detailed mapping and far more powerful computer analysis of spatial relationships.Contents
1 Introduction ... Céline Rozenblat and Guy Melançon
Part I Concepts and Visualizations of Multilevel Spatial Networks
2 A Small World Perspective on Urban Systems ... Céline Rozenblat and Guy Melançon
3 Topological Clustering for Geographical Networks ... Jean-François Gleyze
4 Theoretical Models of Time-Space: The Role of Transport Networks in the Shrinking and Shrivelling of Geographical Space ... Alain L’Hostis
Part II Tools for Networks Analysis
5 Structural Analysis of Networks ... Guy Melançon and Céline Rozenblat
6 Graph Visualization For Geography ... Antoine Lambert, Romain Bourqui, and David Auber
7 Exploring Hierarchies Using the DAGMap ... Pierre-Yves Koenig
Part III Empirical Studies of Spatial Multilevel Networks
8 Ports in a World Maritime System: A Multilevel Analysis ... César Ducruet
9 Comparing Multilevel Clustering Methods on Weighted Graphs: The Case of Worldwide Air Passenger Traffic 2000–2004 ... Céline Rozenblat, Guy Melançon, Romain Bourqui, and David Auber
10 Multilevel Analysis of Corporations Networks: A Comparison Between Agro-Food and Automobile Strategies for Urban Development ... Charles Bohan and Bérengère Gautier
11 The Capture and Diffusion of Knowledge Spillovers: The Influence of the Position of Cities in a Network ... Marie-Noëlle Comin
12 Defining Polycentric Urban Areas Through Commuting Cohesion in France ... Patrice Tissandier, Trung Tien Phan Quang, and Daniel Archambault
Conclusion ... Céline Rozenblat and Guy Melançon
Authors
Terms
Preface
During the last decade, research on networks has developed rapidly in most scientific disciplines. Several factors explain this sudden interest, which has led to an exponential growth in the dedicated literature from the natural to the social sciences. Technical advancements have certainly facilitated this exceptional development. Although graph theory and analytic concepts about graph structure and dynamics did exist before the 1960s (Berge, 1958), the processing of large connectivity matrices remained limited for a long time by the insufficient power of computers. On the empirical side, relational data were scarce or were very expensive to construct. Indeed, the significant increase in the speed of computing that has occurred recently as well as the proliferation of new types of empirical relational data that have been made viable because of the Internet have boosted the creation of powerful tools for network analysis. In parallel, many important modern societal trends have emphasized the need for developing adapted concepts and theories to help understanding these trends, which include the globalization processes that mainly operate through networking relationships; the trend toward decentralized management, the trend toward more participative or cooperative organization than the classical hierarchies; and the unfolding of individual connections through increasing mobility and educational level of population accompanied by various communication tools, to the extent that social networks are now considered to be a major source of “big data”.
Among all of the social sciences, geography is well advanced in analyzing and modeling networks. Physical networks, such as the hydrographic, transportation or infrastructure networks, are part of the structuring of the geographical space. The descriptions of these networks were formalized decades ago, for instance, by models such as Horton’s laws or accessibility and connectivity indices (Garrison, Berry, Marble, Nystuen, & Morrill, 1959). Thereafter, exchanges of goods, people and information between places were considered to be a possible fundamental explanation of geographical diversity leading to a conception of geography as a science of “spatial interaction” (Ullmann, 1954). Theoretical and operational models of spatial interaction were already well established in the 1970s (Wilson, 1970). The gravity model representing trip distribution was even considered to be “the first law of geography” (Tobler, 1970).
However, until recently, there were few methods for fully exploiting the vision of geographical space as a relational space, defined according to the many possible configurations by the interactions between georeferenced entities. Introducing that vision is the most salient aspect of this book. This book develops an integrated set of theoretical interpretations and adapted methods for exploring a variety of networks. These original and reproducible methods were elaborate, due to a long and involved interdisciplinary collaboration, between a laboratory of data processing and a network of geography researchers. A geographer, Céline Rozenblat, and a computer scientist, Guy Melançon, together developed a theoretical conception of geographical networks produced by dynamic processes in complex systems. They also conceived and adapted methods and software for visualizing the specific configurations of relational spaces that are operating at different scales of analysis within these networks. These scholars both succeeded at stimulating interest and animating a group of scientists from Canada, France and Switzerland, who worked together since 2005 when the SPANGEO program was initiated.
A major achievement of this project is the continuous development of the TULIP software, which is dedicated to the visualization of large networks according to the measurements of centralities and proximities in a variety of ways. As the networks analyzed by the geographers connect located objects, they are often rather strongly structured by the constraints that the distance exerts on the practices of communication or displacement. Explaining their structure implies combining the identification of topological organization (for example, small world configuration) and measurement of metrics on weighted interaction flows (representing, for instance, group cohesion). Moreover, as urban centers are the places where social interaction organizes in networks on many scales, from daily commuting patterns to worldwide air transportation system, there was a specific need for integrating within the measurement procedure—the hierarchical structure of urban systems that guides the many patterns of socio-spatial interaction. In addition to the usual methods for detecting communities within the network through various clustering methods, the TULIP software allows for a multi-level analysis of connections that are considered both at the intra-urban and the inter-urban level. This component is especially useful, for instance, in applications to urban economy when exploring the diversity of linkages among firms or among scientific researchers inside—as well as between—cities. This type of tool opens the door to a more specific appraisal of the so-called agglomeration economies, which may in fact represent “network effects”.
Thanks to the efforts of the SPANGEO interdisciplinary group, we warmly welcome the novel feasibility of approaching the geographical space as a relational space. In the contemporary expansion of methods for network analysis, too many workers are satisfied with implementing standard algorithms or measurements on data files that are collected quickly, which yields results that are trivial or are difficult to interpret. This type of data appraisal was not followed in this work. Instead, this work proposes a method that fully integrates geographical theory with topological analysis, providing relevant measurements of centralities and proximities inside a method of visualization. Historically, geographers have used the integrative power of eyesight to envision the peculiarities of landscapes and places, constructing maps for encoding the accumulated knowledge from this process. It is now time to integrate sophisticated visualization tools in the cognitive elaboration of models and theories of spatial interaction that are occurring within the complex geographical systems and, in the process, continuously reshaping them.
Paris, France
Denise Pumain
December 2012
http://link.springer.com/content/pdf/bfm%3A978-94-007-6677-8%2F1.pdf
Sem comentários:
Enviar um comentário