Hierarchical sampling for efficient and comprehensive community connectivity analysis: A Michigan case

Analyzing community connectivity is necessary for identifying construction sites that have various services (e.g., restaurants, supermarkets) within walking distances, so that occupants of the new facilities would not need to use vehicles for accessing those services and thus reduce the carbon footprints of facility operations. Currently, civil engineers need to manually conduct community connectivity analyses on multiple candidate construction sites and recommend the one with the largest number of nearby services. Manual community connectivity analyses are tedious, especially when the engineers need to consider hundreds of locations across an urban area. While experienced engineers may quickly identify some locations with higher community connectivity, such a process is subjective and may compromise the completeness of the solution. To enhance the efficiency of community connectivity analysis while keeping the completeness level of the analysis, this paper explores a hierarchical sampling approach for quickly identifying all locations with high community connectivity across a medium-size city in Michigan. This approach first sparsely sample the urban area (e.g., two miles apart between evaluated locations), and then increase the sampling densities at locations with higher community connectivity based on sparse sampling results. Sensitivity analysis of sampling step sizes are presented and analyzed.