Bypassing the lack of available travel disaggregated data: A geo-spatial framework to simulate mode choice

• Limited access to travel disaggregated data is addressed applying geostatistics. • Conveniently producing disaggregated data utilizing fewer resources. • From traditional spatial models to geostatistical approaches in travel demand. • A pioneering application of Sequential Gaussian Simulation to travel mode choice. • Modeling in non-sampled areas, creating a myriad of spatial scenarios. Disaggregated data for travel demand are essential resources toward appropriate urban planning, especially regarding public transportation. However, especially in developing countries, access to such information is limited. The current paper addresses this issue by introducing an approach, comprising semivariogram deconvolution, Sequential Gaussian Simulation (SGS), and validation, using regular spatial scales. As input to the procedure, we propose to use information with high availability, such as census microdata. The hallmark of the devised method lies in inferring travel spatial variability of more disaggregated unit areas using synthetic data. The method proposes to calculate synthetic data using the socioeconomic census microdata and a calibrated regression model with travel demand data associated with Traffic Analysis Zones (TAZs) in São Paulo city. The resulting maps and statistical metrics corroborate the original data values associated with TAZs. This paper presents relevant contributions as the method enables: bypassing the lack of available travel disaggregated data; creating different scenarios to reproduce travel spatial behavior; and assessing the associated uncertainty.