The implementation of area and membership retrievals in point geography using SQL

Abstract

The Small Area Health Statistics Unit (SAHSU) has been established at the London School of Hygiene and Tropical Medicine (LSHTM) to investigate the geographical distribution of some diseases in the UK. The brief is to hold data on events (including death and cancer registration) identified by geographical location, and to produce event rates and analyses based on arbitrary geographical or administrative aggregations around industrial installations or other points. Events and population counts are associated with very small areas (an irregular tessellation) for which a point coordinate location is available.

Administrative areas form several membership hierarchies, for which the relational model provides a natural retrieval structure. Efficient retrievals for arbitrary geographical areas are achieved through the use of a regular tessellation overlay (grid-squares) as an intermediate step.

Output from the system is as a display of a classified table of rates, or as a file of records (for each aggregate area and sub-group) for further analysis in a statistical package.