Comparison between enriched travel data and the original survey data by means of a model based approach

Abstract

This paper focuses on making a comparison between enriched travel data and the original survey data by means of a model based approach. The research reported in this paper represents a continuation of work undertaken in our previous study ‘The impact of data integration on some important travel indicators’. The Flemish Household Travel (FHTS) data and the Time Use data of the Flemish people (FTUS) data were combined in the previous study using the socio-demographic population data as the base data. The FHTS data and the resultant combined data are then compared here for the duration of travel per person per day by means of the linear regression model. The explanatory variables exploited include sociodemographic variables. The variable relating to the individual trip rates per person per day was also controlled for and the general analyses are conditioned on individual purpose of travel. The results revealed that the combined data set provided a reduction in the magnitude of the standard errors of the parameter estimates indicating higher precision and thus providing a better basis for purposes of prediction. The larger sample (combined data) is therefore invaluable in prediction of travel demand and offers a better base for simulating travel data. Therefore, integrating data from different sources holds out considerable promise for supplementing existent travel survey data making them more enriched for better prediction.