| Abstrakt: |
Bicycling is among the most environmentally sustainable and economically affordable travel modes available. The popularity of bicycling activities strongly depends on the availability of well-connected bicycle networks. Existing methodologies to measure network connectivity are often purely academic, complex, subjective, or locally specific. This study aims to develop and test a reliable methodology for evaluating bicycle network connectivity. The study proposed two weighted shortest-path graph algorithms: the low-stress bike network connectivity (LSBNC), and designated bicycle network connectivity (BNC) algorithms. The weights of the algorithms were the function of slope, level of traffic stress, and link length. We tested the algorithms on the California cities of San Francisco, Davis, Sacramento, and Hayward, along with San Francisco Bay Area counties, and found that algorithms can produce meaningful quantitative connectivity scores. The results indicate that Davis's BNC and LSBNC scores are 0.36 and 0.40, whereas for San Francisco, these scores are 0.07 and 0.47, respectively. The remaining Bay Area county's networks are better connected through a low-stress bike network compared with a designated bicycle network. Finally, we fitted the connectivity scores with the designated bike network or low-stress bike network intersection density and found that the BNC score can be calculated with goodness of fit (R2) of 0.90 and LSBNC can be calculated with R2 of 0.38. The developed methodology will help planners, engineers, and policymakers with the ability to efficiently evaluate bicycle network connectivity. Practical Applications: In general, agencies must understand their network connectivity level before deciding the budget allocation for any infrastructure improvement. This study proposed two novel shortest path–based algorithms that can measure the designated bike or low-stress bike network connectivity for biking. The algorithm used the level of traffic stress, slope, and segment length to calculate the connectivity score. The practitioner can apply our proposed algorithm to obtain the connectivity score at a node or census tract level as well as for the entire network. The connectivity score will range from zero to one, where zero means the network is not connected at all, and one means the network is 100% connected. The obtained connectivity score through our algorithm will help the agency to identify the appropriate portion of the network where improvement is required. [ABSTRACT FROM AUTHOR] |
