Functional regions are formed in the process of urban development and connected by multi-modes public transit. Dynamic regional attributes, which consists of mobility patterns of commuters, not only affect public transportation network planning and station distribution design, but also play an important role in quantitative travel demand forecasting and location portrait shaping. Local Advertisement Campaigns as one applicable example, are in desperate need of targeting strategy covering both regional socioeconomic attributes and dynamic mobility patterns. However, urban residents’ travel patterns depend on their daily activity patterns in the city, and furthermore have close tie with region functional attributes in reverse. In order to help public transportation agency to better understand their customers and accurately retarget them to improve level of service or achieve more income through advertising, we propose an advertising campaign strategy based on region functional attributes and mobility patterns. The strategy framework consists of three steps procedure: First, regional PoIs and bus network data are extracted with both numerical and word expressions and analyzed by topic discovery algorithm to depict Station Portrait (SP), as nearby regional function attributes probability distribution. Second, trip chains are extracted from smart card data, and modeled together with activity imputation by Hidden Markov Model (HMM) for time series mobility patterns discovery. Experiment results show that our strategy considering both mobility patterns and region functional attributes boost accuracy in targeting right group of passengers in public transit, especially in outdoor advertising campaign which rely heavily on travel time and OD.

CICTP