Travel distance computation from household travel survey data: The case of Microcensus 2000

Abstract

Distance related information that has been collected from the travel surveys requires an in-depth analysis for the quality assessment of the directly reported distance data. This paper reports the issues involved in different distances calculations and comparing the same. The objective is to assess the accuracy of the reported travel distances. This report briefly introduced the loaded network distance and comparison of network distance, geo-code based crow-fly distance and the reported travel distance. Data that was collected in 2000 Microcensus is used for comparison. This sample data has been thoroughly screened and only within day car travel was considered for analysis. Variations in trip length distributions of reported distance and the calculated distances are observed. It is found that the reported distances are in between shortest path distance and loaded network distance. Using a descriptive statistical analysis, a series of detour factors are computed for different travel distance ranges. An average detour factor (operational) of 1.61 is obtained form the analysis. Accuracy of the reported information about origin and destination of the long-distance trips is assessed in the end.