This past spring, the City of Boston’s Analytics Team collaborated with the Boston Transportation Department (BTD) to determine where 100 new Bluebike stations (Boston’s bikeshare system) should be placed throughout the city, and why they should go in that location. This project is set to grow Boston’s bikeshare network by 40%- a major impact to the city to expand mobility options for residents.
The Analytics team worked together to apply data science methods and geospatial analysis to make station location recommendations, utilizing data such as job and population density, and ridership and station status. The Bluebikes data was translated into informative metrics such as station trip volumes, instances when stations were empty or full, and a pair of metrics- lost trips and lost returns- that highlighted areas of the city that are not able to meet the existing demand for starting a trip or ending a trip. The team used this data to identify locations with strong latent demand for bike share and to maximize the percentage of households in Boston within close proximity to a Bluebikes station.
In order to provide BTD with specific locations, the team utilized a geospatial suitability modeling approach, inputting dock, population, and job density data, and the new lost trip/return metrics into the model. The suitability modeling was done in two groupings- once for ‘expansion’ stations and one for ‘infill’ stations. The groups had differences in data inputs as well as how the data was weighted and transformed, due to different priorities for the purpose of that station placement. Once the weights and transforms were applied, the team selected the highest scoring locations (with a minimum buffer distance between existing and proposed stations). Currently, BTD is siting the Analytics Team’s proposed station expansion locations, with plans to begin installation within the next year.
