Developing efficient and reliable urban metro systems is regarded as an important and effective approach for mitigating traffic congestion. For many urban metros under development, new lines have been successively built, put into use, and reshaping the travel demands of passengers. In the present study, we employed the online available socioeconomic data, transit infrastructure data and transit service data to develop a machine-learning passenger travel demand prediction model. The model is used to predict passenger travel demand for extending metro networks with new lines successively come into use. The model has also been used to uncover the correlation between each socioeconomic or transportation factor and passenger travel demand, which can provide useful insights for the planning and operation of new metro lines. Keywords: urban metro, new metro lines, OD prediction, travel demand

CICTP