This paper aims to perform a study on the demand for urban public transportation by bus in the context of the city of Salvador. Three econometric models were specified, tested and compared between each other. It was found that the aggregation of variables is one possible way to minimize multicollinearity problems and that improvements in the mathematical modeling enabled a significant improvement in the goodness of fit, maintaining the statistical significance of the parameters. As a result, it was found that the elasticity of the demand in relation to average per capita income was 1.44% (statistically significant at 1%) and, in relation to tariff, it was -0.06% (statistically significant at 10%), revealing an inelastic behavior for this variable. When compared to the work of Carvalho and Pereira (2012), it indicates that the differences in the urban structure of cities should be taken into consideration and that individual studies for each city are necessary. Finally, forecasting in the short term was developed considering evolution scenarios of the analyzed variables.
demand; public transportation; bus; salvador; econometrics