To capture the complex relationships between transportation and land use, researchers and practitioners are increasingly using place-based measures of transportation accessibility to support a broad range of planning goals. This research reviews the state-of-the-art in applied transportation accessibility measurement and performs a comparative evaluation of software tools for calculating accessibility by walking and public transit including ArcGIS Pro, Emme, R5R, and OpenTripPlanner using R and Python, among others. Using a case study of Toronto, we specify both origin-based and regional-scale analysis scenarios and find significant differences in computation time and calculated accessibilities. While the calculated travel time matrices are highly correlated across tools, each tool produces different results for the same origin-destination pair. Comparisons of the estimated accessibilities also reveal evidence of spatial clustering in the ways paths are calculated by some tools relative to others at different locations around the city. With the growing emphasis on accessibility-based planning, analysts should approach the calculation of accessibility with care and recognize the potential for algorithmic dependence in their calculated accessibility results.

中文翻译：

---

计算基于地点的交通可达性：方法、工具和算法依赖性

为了捕捉交通和土地利用之间的复杂关系，研究人员和从业者越来越多地使用基于地点的交通可达性测量来支持广泛的规划目标。本研究回顾了应用交通可达性测量的最新技术，并对用于计算步行和公共交通可达性的软件工具进行了比较评估，包括使用 R 和 Python 的 ArcGIS Pro、Emme、R5R 和 OpenTripPlanner 等。使用多伦多的案例研究，我们指定了基于起源和区域尺度的分析场景，并发现计算时间和计算的可访问性存在显着差异。虽然计算出的旅行时间矩阵在工具之间高度相关，但每个工具对相同的起点-目的地对产生不同的结果。对估计可达性的比较还揭示了空间聚类的证据，这些证据表明，某些工具相对于城市周围不同位置的其他工具计算路径的方式。随着对基于可访问性的规划的日益重视，分析人员应谨慎对待可访问性的计算，并在其计算的可访问性结果中认识到算法依赖的可能性。