Last mile transportation is important in both freight and passenger transport as it accounts for a large portion of the costs and emissions in the transportation industry. In urban transport, the continuously growing travel demands and the rapid development of mass transit systems place a high stress on last mile transportation, which is a vital but underdeveloped part of urban transportation systems. This underdevelopment greatly impedes the further improvement of bus sharing rates and the realisation of sustainable transportation. Therefore, this research proposes a data-driven method to design shuttle services to improve the efficiency and convenience of last mile transportation. Specifically, a unified tool is developed to identify the last mile travel demands from various data sources. Based on these demands, the locations of bus stop are planned through an improved clustering algorithm, and the routing and scheduling of shuttle services are designed using a data-driven method. In addition, a simulation-based cost-benefit analysis is conducted to evaluate the performances of shuttle services in different areas. Finally, a case study using bicycle-sharing data in Shanghai is presented to demonstrate the working process of the proposed method and verify its performance.