Revealing the determinants of gender inequality in urban cycling with large-scale data

EPJ Data Science

Table 2 Street-level indicators for the city of New York: description and source

Variable name	Description	Data source
Unprotected cycleway	Dummy for presence of a shared or unprotected bike-lane	OSM
Protected cycleway	Dummy for either the presence of a protected bike-lane or streets with no vehicles	OSM
Public lighting¹	Dummy for the presence of public lighting	OSM
Unpaved surface	Dummy for unpaved surface	OSM
Park proximity	Dummy for streets next to a park (within 15 meters)	OSM^∗
Any-vehicle crashes	Number of crashes involving any type of vehicles per 10 m of street length	NYC^∗
Bike crashes	Number of bicycle crashes per 10 m of street length	NYC^∗
{borough_name}	Dummy for boroughs (baseline: Manhattan)	NYC
Coast proximity	Dummy for street next to the river coast	OSM^∗
Edge betweenness	Edge betweenness of the streets computed for streets within the largest component of the street network

¹Information on the presence of public lighting is very sparse in OSM. In case of missing information, we assumed that the public lighting is available.
^∗Indicates that the data from the original data sources required specific preprocessing (e.g. for normalization) described in the Methods.