Identifying the Temporal Characteristics of Intra-City Movement Using Taxi Geo-Location Data

In this chapter, the authors focus on temporal patterns of urban taxi trips and explore determinant factors in conjunction with geodatabase at aggregate level. Zero-Inflated Negative Binomial model is proposed in light of count data nature and excessive number of O-D pairs with zero trip. Three typical time slots on weekdays, as well as weekends, are introduced as case study to check temporal variations of intracity movement. The results indicate that trip distance, land use, socioeconomics, and built environment are significant variables that affect the number of taxi trips between two locations. In particular, longer travel and worse economy conditions, such as low employment and average annual income and more population under poverty, may prevent more movements, which have more impacts during peak hours. A better transit system may reduce the taxi trips, except for areas with more subway stations. Develpoed area for instance more commercial or residential area is more likely to attract more visits by taxis, as well as dense public facilities but with more temporal variations. Identifying the Temporal Characteristics of IntraCity Movement Using Taxi Geo-Location Data