Thoughts, Reports and Rambles from the AAG: Urban Applications

Day 2 in Las Vegas and sitting in a session entitled ‘Urban Applications’.

A couple of interesting papers in this session, firstly a paper entitled Who is in your neighborhood? New approaches to define neighborhood context by Iris Hue of UC Berkeley and Weimin Li from Cal Poly Pomona.

Neighborhoods can be defined via various means according to scale and dimension, these factors can be summarized as being: Geo-Spatial, Personal, Social Network, Functional District, Political Community and Economic Entity.

In short, social and behavioral studies have established the importance of neighborhood context in shaping individual behaviors, attitudes, and health outcomes. Although scholars have developed well-theorized definitions of neighborhood, the empirical measurement remains arbitrary and unsatisfactory. On the one hand, Census geographies are often used as proxies for residential neighborhood without considering the exact location of an individual. This approach brings huge uncertainty to the matching of sampled individuals to their neighborhood. On the other hand, many social and behavioral studies employ neighborhood measures at different scale level, e.g., Census block group or tract or city. As a result, these studies sometimes arrive at contradictory results. Better approaches to measure neighborhood context are required to solve these problems.

The paper addresses the limitations and disadvantages of some commonly used measurements, subsequently applying advanced GIS technologies, e.g. geo-processing and geo-statistics, to develop spatially continuous neighborhood from data measured at various scales.

Their approach allows users to explore beyond the conventional definition of residential neighborhood and generate contextual measures beyond what current public data sources can provide. By mapping traffic networks and geo-coding major community amenities and commercial clusters, users can also examine neighborhood context based on activity-space. The approach also takes physical barriers, such as highway or river, as well as varying population density into consideration to develop more precise contextual measures.

Secondly a presentation by Alan G. Phipps from the University of Windsor (that’s not the UK Windsor, in case you were wondering…) with the notable title of: Three Computer-Programmed Applications of Google Maps

Alan notes that computer-programmers have been able for several years to code and display their own data on Google maps via the internet. Google's JavaScript map application is accessible with a personal key for free non-commercial use from its server. He has programmed three different types of Google map applications, and used them in teaching and research:

(1) Point maps of crime and disorder offences, house sales, or exterior house qualities in two Windsor neighbourhoods, for example, at http://web2.uwindsor.ca/courses/sociology/phipps/courses/bquant/uhq2006maps.html#Uhqmap.


(2) Polygon maps of enumeration and dissemination area data from the Canadian census for the same two Windsor neighbourhoods, for example, at http://web2.uwindsor.ca/courses/sociology/phipps/courses/stats/windea01maps.html#Windsormap

(3) Locational maps for automatically geocoding and displaying points of interest, for example, at http://web2.uwindsor.ca/courses/sociology/phipps/courses/plan/haa1.html#AA00.

In demonstrating these types of maps, Alan notes firstly the requirement for pre-analyzed data due to lags in calls to the server from within a programmed application. Secondly, he notes a limitation in the number of displayed data-points before warnings pop up about a slow-running script.

For the last four years attending the AAG we have been amazed at how low profile Google Maps, Google Earth, Microsoft Earth etc systems are. We always feel like going to the AAG we would be at the heart of the it all, instead digital geography, neogeography and Google Maps hacks etc are still reasonably niche topics….