PROPOSAL OF A GEOSTATISTICAL PROCEDURE FOR TRANSPORTATION PLANNING FIELD

Samille Santos Rocha, Anabele Lindner, Cira Souza Pitombo

Resumo


The main objective of this study is to estimate variables related to transportation planning, in particular transit trip production, by proposing a geostatistical procedure. The procedure combines the semivariogram deconvolution and Kriging with External Drift (KED). The method consists of initially assuming a disaggregated systematic sample from aggregate data. Subsequently, KED was applied to estimate the primary variable, considering the population as a secondary input. This research assesses two types of information related to the city of Salvador (Bahia, Brazil): an origin-destination dataset based on a home-interview survey carried out in 1995 and the 2010 census data. Besides standing out for the application of Geostatistics in the field of transportation planning, this paper introduces the concepts of semivariogram deconvolution applied to aggregated travel data. Thus far these aspects have not been explored in the research area. In this way, this paper mainly presents three contributions: 1) estimating urban travel data in unsampled spatial locations; 2) obtaining the values of the variable of interest deriving out of other variables; and 3) introducing a simple semivariogram deconvolution procedure, considering that disaggregated data are not available to maintain the confidentiality of individual data.


Palavras-chave


travel demand; kriging; semivariogram deconvolution; disaggregated systematic sample; aggregated data.

Texto completo:

PDF (English)


Direitos autorais 2017 Samille Santos Rocha, Anabele Lindner, Cira Souza Pitombo

Licença Creative Commons
Esta obra está licenciada sob uma licença Creative Commons Atribuição - NãoComercial 4.0 Internacional.

Boletim de Ciências Geodésicas. ISSN: 1982-2170