Geographically Weighted Regression (GWR) is a method of spatial statistical analysis used to explore geographical differences in the effect of one or more predictor variables upon a response variable. However, as a form of local analysis, it does not scale well to (especially) large data sets because of the repeated processes of fitting and then comparing multiple regression surfaces. A solution is to make use of developing grid infrastructures, such as that provided by the National Grid Service (NGS) in the UK, treating GWR as an “embarrassing parallel” problem and building on existing software platforms to provide a bridge between an open source implementation of GWR (in R) and the grid system. To demonstrate the approach, we apply it to a case study of participation in Higher Education, using GWR to detect spatial variation in social, cultural and demographic indicators of participation.
Grid-enabling Geographically Weighted Regression: A Case Study of Participation in Higher Education in England
Richard Harris, Alex Singleton, Daniel Grose, Chris Brunsdon, Paul Longley
I have blogged about the newly launched London Data Store elsewhere, however I thought I would post a quick (actually not so quick given our ArcGIS liscence server died today) post about a mashup I created today using the CASA software GMapcreator and website Maptube.
Anyway, you can view the map here – it shows deliberate fire incidents in London.
Since Endnote X3 has recently added Pages compatibility I have started to write more on my Macbook. However, one issue with this was that my very large Endnote library was stranded on my university workstation. An effective solution to this problem was to move my Endnote library into my Dropbox folder. For those not familiar with Dropbox, this is free online storage account that comes with an application for Windows / OS X / Linux that syncs files in a specified folder accross multiple machines. This provides a very simple way of sharing Endnote libraries for free!
An alternative, and one which I experiment with in a different context is Zotero . This is a centralized online reference manager with growing caitpabilities that is free to use, and is integrated into Firefox. I am really looking foward to seeing what happens with this application over the next few years. Specifically if they add Pages compatibility.
Happy New Year!
Yesterday I spoke at an OAC User Group event at the RSS about geographic visualisation and area classification more generally. The slides as usual are on slideshare!
Here is my talk from the 39th RSAI:BIS conference on 2-4 September 2009, Limerick, Ireland.
I have just finished creating and adding some content to my new blog Areaprofiles which will provide discussion on the use and creation of geodemographic and neighborhood classifications. You will note an American spelling of neighbourhood in the last sentence and the blog will also have an international flavour. Although area classification feature on other geomarketing blogs, these are predominantly focused on commercial applications. Thus, the main thrust of this blog will be about how area classification have utility for investigation of the form and function of cities rather than how they can be used to sell more units of product X. I also hope to add some tutorials on how classifications can be created from data sources, normalisation procedures and aggregation methods.
I have just finished at the RGS conference which was hosted in the fantastic city of Manchester. I thought the line up this year was very good indeed and there were lots of interesting talks. The UCL talks from CASA comprised the QMRG session I organised on geodemographics and also a postgraduate forum / QMRG badged young researchers session. Some of these talks are online; James Cheshire, Dan Lewis, Oliver O’Brien, Muhammad Adnan and me.
Two events of particular note at the conference were the launch of the International Encyclopedia of Human Geography and the “Mapping Manchester: Cartographic Stories of the City” exhibition organised by Martin Dodge and Chris Perkins (now known as Perkins in the popular press!).
Although viewed by business and commerce as successful solutions, geodemographic profiling of neighbourhoods has attracted wide-ranging criticism in the academic literature. This paper addresses some specific concerns that arise because the derivation of classifications is rarely transparent and open to scrutiny or challenge. The substantive focus of the research reported in this paper is a nationwide geodemographic classification of how people engage with new information and communication technologies (ICTs). In response to the critique of geodemographics as a ‘black box’ technology, we describe how the classification was opened up to public scrutiny and how we conducted a major consultation exercise into the reliability of its results. We assess the message of the 50,000+ searches and 3952 responses collected during the consultation exercise, in terms of possible systematic errors in the shape and detail of the classification. Unusually for Internet-based surveys, we also investigate the likely reliability of the response information received and identify ways in which the outcome of consultation might be used to improve the classification. We believe that this is the first-ever large-scale consultation survey of the validity and remit of a geodemographic classification and that it may have wider implications for the creation of geodemographic classifications.
You can view the article at: International Journal of Geographical Information Science
