This repository contains the sidewalk networks used to produce the results in [1]. The networks for the 10 cities studied are available in both text (edge list and meta-data files) and geographic (ESRI Shapefile) form. Please cite [1] in any works using or derived from the contents of this repository. # Top-level # ## crs_dict.txt ## Each row lists the name of a city and the corresponding meters-based, EPSG registry coordinate reference system (CRS) used in the paper. This can be useful for data visualization or geographic manipulations. Note that the shapefiles in {city}/shp/ are written explicitly using these CRSs. ## ./{city}/ ## The networks for each city are stored in individual directories. Each city directory is further subdivided into a /text/ and /shp/ directory, storing the city's network in text and geographic form respectively. ## ./{city}/text/ ## ### edgelist_length.txt ### Space-separated *Columns* : node\_i node\_j length This format is easily readable by the python network analysis libraries networkx and igraph (also available in R), as well as by Matlab. The third column, length, corresponds to the weight of the edge. Note that while the sidewalk networks are undirected, the edge betweenness calculated for the paper was directed. Thus, each undirected edge appears twice in edgelist.txt, as (i,j) and (j,i) ### edgelist_width.txt ### Space-separated *Columns* : node\_i, node\_j, width Identical format to edgelist.txt above, except that the edge weight here is the width of the sidewalk segment. width\_mat.txt and edgelist.txt can be joined using node\_i and node\_j as an index column. ### node_metadata.txt ### *Semicolon-separated* *Columns:* x, y, population, services Each row contains the metadata for one node of the sidewalk network. Nodes are indexed by row number, starting from 1. These indices correspond to the values of the efirst two columns of edgelist\_length.txt and edgelist\_width.txt, and to the node ids of the shapefiles detailed below in ./{city}/shp/ *x* and *y* : the corresponding coordinates in the coordinate reference system used (see crs\_dict.txt above). *population*: the number of people assigned to that given node, as derived from openly-available governmental population data. *services* encodes the number of Foursquare or OpenStreetMap (in the case of Denver) Points-of-Interest closer to that node than to any other. ## ./{city}/shp/ ## The ./{city}/shp/ directories store the sidewalk networks for each city in geographic (ESRI Shapefile) form. Each city has two shapefiles, representing the network edges and nodes respectively. Note that ESRI Shapefiles (a standard GIS format) are made up of three or more files with the same name but distinct file extensions (i.e. .shp, .shx, .dbf). These files should be kept together in the same directory. For ease of use, the edges and nodes are stored separately in individual directories nested within ./{city}/shp/. These shapefiles are useful for visualising the networks (in a GIS client such as ArcGIS or the open-source QGIS, or using Python with a library like Geopandas), but can also be edited or used as a basis for network analysis. ### edges.shp ### The edges shapefile is made up of polyline geometries. *columns* : uid, to\_from, isCW, ped, length, width *uid* : a unique identifier for the edge. *to_from* : comma-separated list of node ids indicating the nodes that are incident to this edge. These ids correspond to the node ids in all files in ./{city}/text/, as well as the uid field of nodes.shp *isCW* : 0 or 1. 1 indicates that the edge is a crosswalk (i.e., pedestrian crossing, zebra crossing) *ped* : 0 or 1. 1 indicates that the edge is a pedestrian-only or shared-space link, such as a hiking trail, or a "living street" *length* : float, indicates the length of the edge in meters *width* : float, indicates the calculated width of the edge's corresponding sidewalk, in meters. This can be interpreted as an average width (see [1] for more details on width calculation). ### nodes.shp ### The nodes shapefile is made up of point geometries. *columns* : uid *uid*: corresponds both to the node ids indicated in the "to_from" column of edges.shp, and to the node ids of the text-based representation of the network described above (contents of {city}/text/). # References # [1] Rhoads, D., Solé-Ribalta, A., González, M. C., & Borge-Holthoefer, J. A sustainable strategy for Open Streets in (post)pandemic cities. Communications Physics vol. 4 (183), 2021