Date of Award


Degree Type


Degree Name

Bachelor of Applied Sciences Honours


Faculty of Science, Technology and Engineering

First Advisor

Dr L Bloom

Second Advisor

Dr P Pedler


Many researchers today have a need to analyse data in a spatial context. An inherent problem is the mismatch of boundaries between the geographic regions for which data is collected and those regions for which the data is required. Often the solution is to interpolate data from one set of regions to another. This project examines and implements a method of areal interpolation that enables the user to use extra information in areal interpolation to increase the "intelligence ' of the process. This method of Enhanced Areal Interpolation uses a conditional Poisson distribution and the EM algorithm to provide estimated values of a variable. Enhanced Areal Interpolation assumes that data is available for a set of source regions, and is required for a set of target regions. Extra information available about the target regions provides an improved fit of the estimates compared to Areal Weighting Interpolation which uses area proportionality to distribute the data. The theory and concepts are illustrated with an example and implemented using the software packages Maplnfo version 3 for Windows and MapBasic version 3 for Windows.