In the UK, gridded census outputs currently utilise a uniform 1km2 grid. While such grids provide a more uniform resolution than traditional census geographies like output areas, they are of limited use in densely-populated areas. To address this, ONS Geospatial have developed the nested grid: this has variably-sized cells, based upon the data within them and the relevant disclosure limits. This will allow data to be disseminated at the most granular resolution and gives users the ability to create and develop insights into the data at a previously unavailable level of detail.
Initial research produced a set of perfectly nested grids, with cell sizes of 1000m, 500m, 250m, and 125m. Each cell was given a unique ID which specified the cell size and its position within the hierarchy.
Subsequent work has focused on creating a methodology for dynamically populating the grids using R. We have created a set of rules based on the underlying data that govern cell behaviour and which also allow for data perturbation in certain settings. These rules ensure that the smallest possible cell size is used, without breaching the disclosure limits. In this presentation we will explain the grid nesting methodology and present a case study of Eastleigh, UK, where we used synthetic household data to test and develop the process.
Start date: Thursday, 10 October 2019 | End date: Friday, 11 October 2019
All content © Delegate Live Limited. All rights reserved.