Estimating Population Trends for Census Tracts: Street Weighted Areal Interpolation Using GIS

Michael Reibel, California State Polytechnic University

Small area demography often requires combinations of data spatially aggregated to incompatible zone systems in a region, such as a set of enumeration districts that changes over time. Such situations can be addressed by areal interpolation, the transfer of data between zonal systems according to spatial algorithms. The authors test a technique of areal interpolation using Geographic Information Systems (GIS) that uses a digital map layer representing streets and roads to derive weights for small areas within aggregation zones. Errors in estimation are measured by comparing local estimates to true local source counts aggregated to target areas from smaller zones. The street weighting technique reduces errors in estimation as compared to estimates derived using the commonly applied area weighting technique, and it is much easier to use than other interpolation techniques that have also been shown to reduce error. Keywords: areal interpolation, dasymetric mapping, small area demography, census geography