The sensor-era has brought rapid changes in transportation; the abundance of data has started changing the traditional way in which planners and engineers approach mobility. Nowadays, traffic monitoring and information provision systems heavily rely on floating car data usually of special vehicles (e.g., trucks, taxi), and the question that arises is whether such sources can provide reliable data for the whole traffic in a complex urban environment. The current paper, through Thessaloniki's (GR) case study, seeks to evaluate the reliability of taxi data compared to the overall traffic. The analysis reveals that for the examined critical urban road paths, there is a strong relation among floating taxi data with the overall traffic that is additionally influenced by other significant factors (e.g., number of lanes, day, time period). Furthermore, a modelling approach with a generalized linear model (gamma with log link) seems appropriate when dealing with skewed and heteroscedastic traffic data.
A Bayesian approach for Generalized Linear Model Using Non-local Prior (Case Study: Poverty Status in East Java)
