Data Collection and Econometric Analysis of the Demand for Non-motorized Transportation

摘要: In this project, a latent class model was derived with assignment mechanism based on the bicycle status of respondent. Two segments were identified: more-skilled and experienced cyclists, versus less-skilled- non-cyclists. The two have different sensitivities to factors that may encourage or discourage riding bike. For instance, slope inclination is considered almost 3 times as bad by less-skilled cyclists. Heavy traffic affects twice much who also consider rain be 2.4 more bothersome (and snow 4 bothersome) than On other hand, bike lanes are 1.6 appreciated Because in cycling route decisions there no direct monetary cost involved, analyze differences taste parameters authors proposed use ratio marginal rate substitution respect travel time. addition, diminishing negative effect hilly topography (slope inclination) measured function physical condition cyclist. terms policy recommendations, results suggest provision an increase modal share cycling, especially among those individuals using infrequently, mostly for recreational purposes. This project examined performance several ridership prediction models, including Negative Binomial regression time-series models such SARIMA SARIMAX. Using counts Portland, show SARIMAX includes weather conditions (temperature precipitation) explanatory variables performs best out-of-sample prediction. Future research State Space needed overcome problems when predicting periods really poor weather. sum, both discrete choice time series analyses coincide indeed main determinant discouraging transportation alternative.