Optimal station localization for bicycle sharing system

Many urban planning strategies for sustainable cities (New Urbanism, Mixed Use Development and Compact City) tout the need for smaller more compact communities [1, 2, 3]. These urban forms depend on a modal shift from automobiles to walking and cycling. As this shift is unlikely to be driven by the markets, increases in sustainable modal transport, infrastructure and disincentives for driving [4] are necessary [5]. In some cities, bicycle sharing systems (BSS) have been catalysts for increasing cycling through their ability to extend existing public transit systems, encourage cycling accessibility and visibility. Cycling has many direct and indirect health benefits such as cardiovascular and reduced CO emissions, compared to car transport, respectively as well as social and equity benefits [6, 5:123]. BSSs are composed of bicycles and spatially fixed stations containing a constant number of docks. Two important advantages of the current generation BSSs are simplicity and security [7]; bicycle theft is difficult yet subscribers may checkout bicycles quickly and without difficulty, allowing users to seamlessly transfer between transport modes. The central question of this research is determining whether a mismatch occurs between BSS’s stations’ optimal locations based on origin and destination demands of potential cyclists and the existing stations. Most BSSs are being managed by private corporations, which often use advertising to subsidize the systems. BSS stations, or kiosks, often have double purposes of providing BSS services, such as checking out a bike, and displaying advertising. Effectively locating BSS stations throughout an urban fabric is of crucial importance to the adoption of the BSS. The optimal locations for advertising billboards and BSS stations are unlikely to consistently correspond. This conflict of interest has been explicitly anticipated by at least one city developing a BSS [8]. The overlap between optimal advertiser locations and stations for BSSs are destinations of high traffic such as commercial and transportation hubs. Conversely, BSS stations may insufficiently serve residential areas far from or without through-roads due to having decreased traffic and therefore marketing potential. This project addresses the question in three steps. The first is the gathering of spatial and spatio-temporal data related to cycling demand in order to develop demand surfaces. Next, demand is contrasted against existing BSS station infrastructure. Assessment is complex due to the alternate realities between data, theory and the practical operations of placing stations within available space constraints and the varying BSS priorities, such as social or racial equity, of cities. Finally we develop a model that optimally locates BSS stations for public use across the urban fabric based on demand and network distances between stations. Performing this process across multiple cities with BSSs we bring insight into whether a localization bias exists towards high throughput and advertising valued locations. Bicycle sharing systems may become the first new ubiquitous ‘public’ transport system in over a century. However, the initial privatization of this new transport system, whose impacts and evolution are not yet completely understood, may result in inefficiencies with lasting negative outcomes. The inadequate development of a city’s BSS may cause it to fail and stigmatize a progressive sustainable transport alternative for a generation