Level of traffic stress-based classification: A clustering approach for Bogotá, Colombia

Abstract

Highlights Data-informed methodology calculates the level of traffic stress of cyclists. Method scales to massive data sets by coupling a classifier with a predictive model. Methodology tested on the road network of Bogotá (Colombia) Web-enabled dashboard supports policy making and interventions to reduce stress. Number of bicyclists’ collisions per kilometer correlates with higher stress.

Is the built-environment at origin, on route, and at destination associated with bicycle commuting? A gender-informed approach

“…Five secondary data sources were used for the generation of community, built-environment and natural variables, including 1) the local geospatial office IDECA ( Unidad Administrativa Especial de Catastro Distrital, 2018 ), 2) the official record of thefts and robberies of Bogotá ( Policía Nacional de Colombia, 2018 ), 3) the compilation of collisions used in Carvajal et al ( Carvajal et al, 2020 ), 4) the Japanese Aerospace Exploration Agency (JAXA) ( Tadono et al, 2016 ), and 5) the level of traffic stress-based classification of the roads of Bogotá ( Huertas et al, 2020 ) .…”

“…Road network characteristics included the proportion of roads classified at different Levels of Traffic Stress (LTS). The proportion of roads classified as low LTS and extremely high LTS were calculated using available road network data for Bogotá ( Huertas et al, 2020 ). Road segments labeled with low LTS are characterized by narrow streets, few lanes, low speed, low traffic density, low traffic flow, and low congestion.…”

“…In contrast, increased public transport accessibility diminishes the likelihood of bicycling ( Cole-Hunter et al, 2015 ; Dickinson et al, 2003 ; Parkin et al, 2008 ; Stinson and Bhat, 2004 ; Timperio et al, 2006 ). As well as road characteristics that create stressful conditions for the bicyclists like those summarized by the Level of Traffic Stress (LTS) ( Furth et al, 2016 ; Huertas et al, 2020 ; Mekuria et al, 2012 ).…”

Bicycle safety in Bogotá: A seven-year analysis of bicyclists’ collisions and fatalities

“…Finally, we include the input variables that allow us to compute the LTS for road segments and intersections ( Huertas et al, 2020 ) at the crash site. These measurements (input variables) comprise lane width, average motor vehicles’ speed, average congestion, vehicles’ density, average vehicles’ flow, number of lanes, the presence/absence of bicycle infrastructure, and the presence/absence of public transportation system (SITP, Sistema Integrado de Transporte Público) routes (a proxy for the presence of heavy vehicles).…”

“…To estimate the values of the average motor vehicles’ speed, average congestion, vehicles’ density, and the average vehicles’ flow, we retrieved the average and free-flow transit times for every road segment using Google’s distance matrix API. We use the traversing time along the segment at 24:00 of Tuesday, September 4, 2018, as a proxy for the free-flow traversing time ( Huertas et al, 2020 ). We included the LTS rating from 1 to 4, as shown in Appendix Table 3, and also considered only differentiating LTS level 4 (higher stress = 1) from all other levels (levels 1, 2, and 3 = 0).…”