Algorithms for Reengineering 1991 Census Geography

Abstract

"The availability of GIS [geographic information systems] technology and digital boundaries of census output areas now makes it possible for users to design their own census geography. Three algorithms are described that can be used for this purpose. An Arc/Info implementation is briefly outlined and case studies presented to demonstrate some of the results of explicitly designing zoning systems for use with 1991 [U.K.] census data."

Extending the automated zoning procedure to reconcile incompatible zoning systems

“…Application of these zone design methodologies to published census data is demonstrated by Openshaw and Rao (1995), who consider the 'reengineering' of 1991 UK Census outputs by using enumeration districts (EDs) as input building blocks, assembling larger zones using a variety of objective criteria. Openshaw and Alvanides (1999) demonstrate similar applications using a national application with ward-level data.…”

“…In both cases, the census user is encouraged to reaggregate the standard zones for which data are published to provide larger, purpose-specific zones designed according to clearly defined criteria. Openshaw and Rao (1995) discuss various search algorithms including the simple AZP, simulated annealing (SA) in which suboptimal swaps are permitted in the early stages of iteration, allowing the procedure to escape from potential local suboptima and a tabu search approach, in which recently tried swaps cannot be considered again until a certain number of iterations have elapsed. There has also been consideration of parallel implementations of AZP, although the partitioning of the problem into independent spatial sub-regions is problematic due to the fact that a boundary reconfiguration may have ramifications across the whole of the problem space.…”

Simulation based population synthesis

“…For each zone the population is synthesized by replicating the sample and optimizing these weights in order to minimize the dierence from the zonal marginals (Voas and Williamson, 2000;Lu, 2011). Openshaw and Rao (1995), Williamson et al (1998), and Voas and Williamson (2000) used Simulated Annealing (SA) as an optimization tool to produce the synthetic population. Ryan et al (2009) compared the CO based methods with the IPF and concluded that CO produced lesser variance.…”

Distribuição da população e cobertura da terra: o lugar das Áreas Protegidas no Pará, Brasil em 2010

“…6 A transferência dos dados do Universo do Censo Demográfico 2010 (população residente, domicílios ocupados e situação do domicílio) para a grade estatística se deu por 5 Qualquer análise espacial está sujeita ao MAUP, mas os efeitos de escala e de zoneamento ficam mais críticos à medida que a quantidade de unidades diminui e a dimensão das unidades aumenta (WRIGLEY, 1995). Ainda que não seja unânime a afirmação de que a agregação de dados em níveis geográficos menores minimize os impactos do MAUP (OPENSHAW; RAO, 1995), a estratégia é defendida por Goodchild (1992) e Schuurman et al (2006), entre outros. Como um sistema de grade estatística apresenta uma grande quantidade de células pequenas, ao contrário de um sistema de unidades administrativas com poucas unidades de grande extensão, infere-se que o MAUP é menos crítico no primeiro caso.…”