Dear David and other respondents, Thanks for the responses to this query. Having just returned from Liverpool, where I presented my results of nearest neighbour clusters, I received a fairly positive response. Last year I did some analysis of 'sustainable' housing areas based on four domains of indicators: cohesion (crime, voting and population change), exclusion (broadly poverty related indicators), housing popularity (turnover, voids and house prices) and environment and infrastructure (derelict land and community and retail facilities). The approach was largely validated by the literature in these areas and pragmatic (what data was available). The result was that there was not a clear relationship between all of the domains, although in the inner part of Liverpool there was a large degree of overlap. Focus groups and interviews with key actors plus around 10 seminars with policy wonks confirmed the broad spatial patterns emerging. For this stage I wanted to develop the methodology to focus on the inner part of Liverpool but to create clusrters of neighbourhoods that incorporated the adjacent EDs that did not meet the threshold of being in the highest quartile on each domain. The result is a more strategic focus on four areas. I guess that David's suggestion of using cluster analysis is partly addressed by my use of domains, but rather than letting the statistics do the work of combining indicators I thought that it best to look at 'meaningful' domains. Regards Peter > This isn't a statistical methodological answer to Peter's question - more a report of long term practice. My approach is to avoid > uni-dimensional index construction and instead classify EDs using sets of relevant indices as input to cluster analysis procedures. > The really significant alogarithm choice is of the set of variables used to construct the clusters. I then map by colouring in a > physical map with coloured pencils with a different colour for each cluster. I am sure there is a much more high tech method of > mapping but I actually rather enjoy doing this. My experience, and I have done this a lot over the years, is that you don't get > 'chaotic' clusters. I am assuming Peter is using Census data so there is no real (10% sample is very big and most indices I use come > from the 100% set) sampling issue. If you find something odd it is really there but in general what you get is adjacent areas of EDs > constituting neighbourhoods which triangulate very well with administrative data and qualitative reporting. > > An alternative would be to use Census Tracts - a very handy size between Wards - far too big - and EDs which are small. > > > David Byrne > > > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Peter Lee Centre for Urban and Regional Studies (CURS) The University of Birmingham Pritchatts Road Edgbaston Birmingham B15 2TT UK Tel (+44) 121 414 3645 Fax (+44) 121 414 4989 http://www.bham.ac.uk/curs/ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%