Feasibility of estimating travel demand using geolocations of social media data

Abstract Travel demand estimation, as represented by an origin–destination (OD) matrix, is essential for urban planning and management. Compared to data typically used in travel the key strengths of social media are that they low-cost, abundant, available real-time, free geographical partition. However, also have significant limitations: population behavioural biases, lack important information such trip purpose demographics. This study systematically explores feasibility using geolocations Twitter estimation examining effects sparsity, spatial scale, sampling methods, sample size. We show suitable modelling overall average weekday but not commuting demand, due low reliability identifying home workplace. Collecting more detailed, long-term individual from user timelines a small number individuals produces accurate results than short-term much larger within region. developed novel approach geotagged tweets attraction generators opposed commonly adopted generators. significantly increases usable data, resulting better representation demand. demonstrates can be viable option estimating though careful consideration must given method, model,