Understanding spatial-temporal travel demand of private and shared e-bikes as a feeder mode of metro stations

Private e-bikes and shared e-bikes are gradually becoming the preferred modes of feeders for a vast number of metro passengers in China, opening up new potential for sustainable urban transportation development. This study proposes a method to identify feeder behaviors from private and shared e-bikes to the metro using mobile phone signal data and shared e-bike operation order data in Nanning, China. Then, the Light Gradient Boosting Machine models are constructed to reveal the influence of various factors on the feeder demand for two kinds of e-bikes in metro stations, to promote integrated travel of e-bikes to the metro. The results show that the demand for private e-bikes in metro stations is significantly higher than that of shared e-bikes, and the demand for the two types of feeder modes during the morning and evening peak hours on weekdays is greater than that in other periods. Secondly, the density of educational facilities has a positive effect on the demand for both feeder modes, and it has a greater impact on the demand for private e-bikes. Thirdly, the distance to the city center has a non-linear effect on the demand for private e-bikes. The farther away from the city center, the more feeders use private e-bikes to travel, and the fewer feeders use shared e-bikes. These findings can help planners better un-derstand how various factors influence feeder demand.