Calculating accessibility with impedance

Data sources

• Roads – US census bureau TIGER Polyline shape files
• Speed limits – Census.gov for classification, exterior sources for the actual figures.
• Crime data - 2016 crime data by AGS polygon shape files, ArcGIS online resources
• Light at night – NOAA radiance Rasters.
• Facilities, landmarks, parks and other geographic data, point shape files – ESRI maps and data 2016

Destinations

Origins

Road network

Splitting lines at intersections & adding new nodes

The streets layer was pre-processed using the Planerize Lines tool in ArcGIS in order to split lines at all intersections.

The roads were further split to allow origin and destination nodes to connect with the network, using the Split Line at Point tool in ArcGIS.

To find where to split, we used the snapPointsToLines function (package maptools) in R, for example:

snapPointsToLines(
  points = orig, 
  lines = sl, 
  withAttrs = FALSE
)

Origin and destination points that are >280 m from the nearest road were removed.

Lines to network

The processed line layer was converted to a “network” object (class SpatialLinesNetwork, package stplanr).

sl = as(sl, "Spatial")
sln = SpatialLinesNetwork(sl)

# Define weights
sln@sl$length_km = SpatialLinesLengths(sln@sl)
sln@sl$time_h = sln@sl$length_km / sln@sl$speed_kmh
E(sln@g)$weight = sln@sl$time_h
sln@weightfield = "time_h"

sln@sl

## Simple feature collection with 291196 features and 15 fields
## geometry type:  LINESTRING
## dimension:      XY
## bbox:           xmin: -71.61874 ymin: 41.89704 xmax: -70.63903 ymax: 42.7226
## epsg (SRID):    4326
## proj4string:    +proj=longlat +datum=WGS84 +no_defs
## First 10 features:
##    OBJECTID              FULLNAME RTTYP MTFCC Shape_Leng speed_kmh
## 1    103007           State Rte 9     S S1200 0.26876997       104
## 2    111017 Boston Worcester Tpke     M S1200 0.26763102       104
## 3    111536           Turnpike Rd     M S1200 0.16620237       104
## 4     92620           Hundreds Rd     M S1400 0.01432285        88
## 5     92620           Hundreds Rd     M S1400 0.01432285        88
## 6    103007           State Rte 9     S S1200 0.26876997       104
## 7    111017 Boston Worcester Tpke     M S1200 0.26763102       104
## 8    111536           Turnpike Rd     M S1200 0.16620237       104
## 9     99017           State Rte 9     S S1200 0.26885250       104
## 10   103208           Turnpike Rd     M S1200 0.16651601       104
##                          geometry id      crime        lan         dens
## 1  LINESTRING (-71.59387 42.28...  1 0.03367609 0.01169225 0.0001769729
## 2  LINESTRING (-71.59387 42.28...  2 0.03367609 0.01169225 0.0001769729
## 3  LINESTRING (-71.59387 42.28...  3 0.03367609 0.01169225 0.0001769729
## 4  LINESTRING (-71.58981 42.28...  4 0.03367609 0.01169225 0.0001769729
## 5  LINESTRING (-71.59143 42.28...  5 0.03367609 0.01169225 0.0001769729
## 6  LINESTRING (-71.59116 42.28...  6 0.03367609 0.01169225 0.0001769729
## 7  LINESTRING (-71.59116 42.28...  7 0.03367609 0.01169225 0.0001769729
## 8  LINESTRING (-71.59116 42.28...  8 0.03367609 0.01169225 0.0001769729
## 9  LINESTRING (-71.59197 42.28...  9 0.03367609 0.01169225 0.0001769729
## 10 LINESTRING (-71.59197 42.28... 10 0.03367609 0.01169225 0.0001769729
##    speed_kmh1       length length_km      time_h     time_h1
## 1   101.69527 237.2447 [m] 0.2372447 0.002281199 0.002332898
## 2   101.69527 237.2447 [m] 0.2372447 0.002281199 0.002332898
## 3   101.69527 237.2447 [m] 0.2372447 0.002281199 0.002332898
## 4    86.04985 152.6427 [m] 0.1526427 0.001734576 0.001773887
## 5    86.04985 262.3491 [m] 0.2623491 0.002981239 0.003048803
## 6   101.69527 178.4970 [m] 0.1784970 0.001716318 0.001755215
## 7   101.69527 178.4970 [m] 0.1784970 0.001716318 0.001755215
## 8   101.69527 178.4970 [m] 0.1784970 0.001716318 0.001755215
## 9   101.69527 527.0206 [m] 0.5270206 0.005067506 0.005182351
## 10  101.69527 527.0206 [m] 0.5270206 0.005067506 0.005182351

sln@g

## IGRAPH 89ac23c U-W- 209873 291196 -- 
## + attr: x (g/n), y (g/n), n (g/n), weight (e/n)
## + edges from 89ac23c:
##  [1]  1-- 2  1-- 2  1-- 2  3-- 4  2-- 4  2-- 5  2-- 5  2-- 5  6-- 7  6-- 7
## [11]  6-- 7  8-- 9  9--10 11--12 10--12  5--13  5--13  5--13 13--14 12--14
## [21]  7--13 12--15 15--16 16--17 17--18 17--19 18--20 20--21 20--21 15--22
## [31] 16--22 19--23 23--24 18--24 13--25 13--25 13--25 22--25 25--26 25--26
## [41] 25--26 23--26  7--27  7--27  7--27 27--28 27--28 27--28 28--29 30--31
## [51] 20--31 26--32 26--32 26--32 30--33 33--34 30--35 35--36 35--37 32--38
## [61] 32--39 32--39 32--39 28--40 28--40 28--40 40--41 42--43 38--42 40--44
## [71] 40--44 40--44 45--46 44--47 48--49 47--50 51--52 53--54 51--55 39--55
## + ... omitted several edges

Impedance

# Rescale
sl$crime = rescale(sl$crime, c(0, 0.1))
sl$lan = rescale(sl$lan, c(0, 0.1))
sl$dens = rescale(sl$dens, c(0, 0.1))

# 'speed_kmh' reduction
sl$speed_kmh1 = sl$speed_kmh + 
  (-sl$speed_kmh * sl$crime) + 
  (sl$speed_kmh * sl$lan) + 
  (-sl$speed_kmh * sl$dens)

Finding nearest nodes

orig_nodes = find_network_nodes(sln, coordinates(orig), maxdist = 1000)
dest_nodes = find_network_nodes(sln, coordinates(dest), maxdist = 1000)