Jean-Claude Thill

Social area analysis, data mining, and GIS

There is a long cartographic tradition of describing cities through a focus on the characteristics of their residents. A review of the history of this type of urban social analysis highlights some persistent challenges. In this paper existing geodemographic approaches are extended through coupling the Kohonen Self-Organizing Map algorithm (SOM), a data-mining technique, with geographic information systems (GIS). This approach allows the construction of linked maps of social (attribute) and geographic space. This novel type of geodemographic classification allows ad hoc hierarchical groupings and exploration of the relationship between social similarity and geographic proximity. It allows one to filter complex demographic datasets and is capable of highlighting general social patterns while retaining the fundamental social fingerprints of a city. A dataset describing 79 attributes of the 2217 census tracts in New York City is analyzed to illustrate the technique. Pairs of social and geographic maps are formally compared using simple pattern metrics. Our analysis of New York City calls into question some assumptions about the functional form of spatial relationships that underlie many modeling and statistical techniques.

Local Indicators of Network-Constrained Clusters in Spatial Patterns Represented by a Link Attribute

Clustering in a spatially distributed phenomenon is an important focus of spatial analysis because it not only suggests characteristics of the pattern itself but also of its background processes. Traditional methods of spatial cluster detection mostly rely on the planar space assumption, yet a variety of spatial phenomena do not support its logic. This article expounds on an exploratory spatial data analysis methodology named local indicators of network-constrained clusters (LINCS) introduced elsewhere for detecting local-scale clustering in a spatial phenomenon that is constrained by a network space. In particular, this article focuses on two types of LINCS methods that are network extensions of traditional methods for analyzing spatial associations in zone-based planar-space data, namely, the local Moran I statistic and the local Getis and Ord G statistic. They are designed for phenomena that are represented by attribute values of individual network links. Examples of such phenomena include traffic volume, traffic speed, and the number of vehicle crashes aggregated at the link level. When the phenomenon of interest can be seen as a subset of a more generic spatial phenomenon, for example, vehicle crashes in relation to the entire traffic observed in a study region, the LINCS methods are capable of taking into account the distribution of such a base phenomenon so that one can avoid the detection of spurious clusters merely reflecting the base distribution. The article illustrates the application of the two LINCS methods using data on highway vehicle crashes in the Buffalo, New York, area in 1997.

Visual Data Mining in Spatial Interaction Analysis with Self-Organizing Maps:

Given that many spatial interaction (SI) systems are often constituted in large databases with high thematic dimensionality, data complexity reduction tasks are essential. The opportunity exists for researchers to examine the formation of different types of SIs as well as their interdependencies by exploring the patterns embedded in the data. To circumvent the limitations of existing methods of flow data compression and visual exploration, we propose an integrated computational and visual approach, known as VISIDAMIN, for handling both SI data projection and SI data quantization at once. The computational method of self-organizing maps serves as the data mining engine in this process. Using a large domestic air travel dataset as a case study, we examine how the characteristics of the air transport system interact with the SI system to create relationships and structures within the US domestic airline market.

Analysis of traffic hazard intensity: A spatial epidemiology case study of urban pedestrians

Traffic safety studies have underscored the hazardous conditions of pedestrians in the United States. This situation calls for increased public awareness of the pedestrian safety issue and better knowledge of the main factors contributing to traffic hazard for urban pedestrians. The purpose of this spatial epidemiology research is to gain greater insights into the geographic dimension exhibited by the intensity of traffic collisions involving urban pedestrians. Pedestrian crashes are studied in Buffalo, NY for years 2003 and 2004. Factors of hazard intensity are determined and compared for three age cohorts as well as for collisions occurring at intersections versus mid-block locations. Physical road characteristics and density of development, as well as socio-economic and demographic variables and potential trip attractors are examined. Spatial regression models are used to account for spatial dependencies. Econometric analysis underscores that all classes of environmental factors tested are significant drivers of pedestrian traffic hazard intensity. Results of the geographic analysis indicate that young and adult pedestrian traffic hazard intensities follow rather distinct logics. In addition, intersection and mid-block crashes differ by their socio-economic correlates, as well as their spatial distribution in the urban fabric.

Delineating Urban Housing Submarkets with Fuzzy Clustering

It has long been argued that the housing market is spatially compartmentalized within a metropolitan area. The argument has important implications for explaining how the housing market works—should the status quo be seen as an equilibrium state? Or if no equilibrium is reached, how do loosely interlaced submarkets function both independently and interdependently? We note that the body of literature has leaned toward testing the distinctiveness of housing submarkets given a priori housing submarkets. However, there seems to be a lack of interest in developing methods for deriving housing submarkets empirically. Fuzzy clustering is well suited to this problem, given that the boundary of housing submarkets is not often sharply delineated. The study applies a fuzzy c-means (FCM) algorithm to identify housing submarkets in the Buffalo–Niagara Falls region. The study is distinct from other FCM applications in three respects. First, we reflect on issues tied to choosing parameters of fuzzy clustering. Second, we introduce overlap measures to characterize the relationship between the clusters produced. Third, we evaluate the performance of fuzzy clustering in terms of hedonic prediction accuracy. Results show that stratified hedonic models predict house price better than a market-wide hedonic model. Fuzzy clustering solutions also yield better prediction, compared with hard clustering.

Urban Bicyclists: Spatial Analysis of Adult and Youth Traffic Hazard Intensity

As issues related to oil dependency, rising gas prices, and global warming come to the forefront of topics of concern for Americans, the need for alternative modes of transportation has become critical. Urban settings are seemingly ideal for bicycling to become a significant mode, given the greater compactness of destinations. However, in the United States, bicycling is both scarcely used and very dangerous, as bicyclists are 12 times more likely to be killed than automobile drivers. The purpose of this research is to gain greater insights into the geographic dimensions of traffic crash intensity that bicyclists may experience in American cities. Bicycle crashes are studied in Buffalo, New York, for the years 2003 and 2004. The geographic distribution of crashes is determined and compared for both youth and adult bicyclists and factors of crash hazard intensity are statistically identified. Density of development and physical road characteristics such as roadway and intersection functional class are examined, as well as socioeconomic and demographic variables and potential trip attractors. Given the spatial nature of these variables, a spatially weighted regression model is incorporated to account for spatial dependencies of the dependent variables and of their model residuals. The results of the analysis indicate clear distinctions between youth and adult bicycle crashes, both in terms of the neighborhoods where victims reside and in terms of the neighborhoods where these two demographic groups are found to be more frequently involved in a crash with a motorized vehicle.

Tree Induction of Spatial Choice Behavior

Discussed is the merit of inductive learning as a tool with which to discover geographic knowledge in data-rich environments, and particularly as an analysis tool in spatial decision-making theory. The capability and applicability of Quinlan’s C4.5 decision tree induction algorithm to the class of problems involving the choice among discrete travel destinations within an urban area are analyzed. The C4.5 algorithm and its relation to other decision tree induction algorithms and to spatial behavioral modeling are described, and its implementation on spatial behavior data from the Minneapolis-St. Paul metropolitan area is illustrated.

Trajectories of Multidimensional Neighbourhood Quality of Life Change

This paper provides an empirical analysis of the multidimensional, spatio-temporal quality of life (QoL) trends followed by neighbourhoods in Charlotte, NC, between 2000 and 2010. Employing a combined geocomputational and visual technique based on the self-organising map, the study addresses which types of neighbourhood experienced the most change or stability, where (in attribute and geographical spaces) did neighbourhoods that began the decade with a particular set of characteristics evolve to, and where did neighbourhoods that concluded the decade transition from? Results indicate that the highest QoL neighbourhoods were most stable, while those with lower homeownership, closer to the city centre, exhibited the sharpest longitudinal trajectories. Lower-income neighbourhoods are found to be heterogeneous in terms of their social problems, dividing between high crime concentrations and youth-related social problems. An exchange of these social issues over time is observed as well as a geographical spread of crime to middle-ring suburbs.

Functionally critical locations in an urban transportation network: identification and space-time analysis using taxi trajectories

This paper studies the space-time properties of locations that are critical to travel activities in an urban environment. Specifically, we analyze locations on the urban street network from the perspective of the distribution of peoples travel trajectories. We identify the intersections of an urban transportation network which are characterized by good connectivity, serving a high density of trip trajectories, and exhibiting multiple traversing patterns of trip trajectories as potential functionally critical network locations (FCNLs). A geospatial method is proposed to extract FCNLs from peoples moving trajectories based on the street network. Two groups of quantitative indices are introduced to measure the evolution of the spatial extent and temporal variation patterns of different criticality levels of FCNLs. A case study using taxi trajectory data from Wuhan, China has been implemented. The results show that the FCNLs are very powerful in uncovering the space-time traveling patterns of a particular population and studying the relationship between urban functional structures and peoples activities. Language: en

Intermodal Freight Transportation and Regional Accessibility in the United States

The authors investigate how the intermodal freight-transportation network affects the ability of regions to position themselves more effectively in the national space economy. The case of domestic containerized freight traffic is examined because it is closely associated with contemporary forms of integration between rail shipping and trucking. With the help of a geographic information system, the potential impact of intermodalism in the United States is analyzed by mapping integral place accessibility measures of five-digit zip-code areas. The performance of the intermodal freight network is evaluated by comparing accessibility measures based on the highway network and on the intermodal network, respectively. Geographically weighted regressions are also performed to identify the variables that contribute to the improvement of accessibility due to intermodalism, while accounting for the spatial nonstationarity of relationships.