Introduction

As part of helping to draft a new policy we need to consider areas within 400m of secondary schools. Buffers have previously been used for such measurements i.e. as-the-crow-flies. However, for the draft policy a more accurate measurement would be a 400m walk and using isochrones. The diagrams below illustrate the difference between a buffer and an isochrone:

Buffer

Isochrone

Strictly speaking, a 400m distance from a school is an isodistance and a 5 minute walk is an isochrone. However, the term isochrone is generally used to cover both distance and time from a point.

The purpose of these notes is to understand:

The best network dataset/service and software to use to create the isochrones.
How to define the school points e.g. the main school entrance?

The best network dataset/service and software to use will be looked at in descending order of convenience i.e. those further down the list require more effort:

AGOL logistic service and ArcGIS Pro
openrouteservice and R
OS MasterMap Highways Network and Portal
OS MasterMap Highways Network and ArcGIS Pro

The resulting isochrones will be included in the Portal App (requires a license and login credentials), which was created to help draft the policy.

Secondary schools

The first thing to do, is get the secondary school features we want to create isochrones for. I’ve made some notes about the Council’s GIS, and more specifically about the R-ArcGIS Bridge that is used below to get the data from the Council’s web GIS.

# URL for secondary schools from the Council's Portal (without a token!)
secondary_schools_url  <- str_c(
  "https://sheffieldcitycouncil.cloud.esriuk.com",
  "/server/rest/services/AGOL/Education/FeatureServer/3")

# Get secondary schools as simple features with WSG84 coordinates
sf_secondary_schools <- arc.open(secondary_schools_url) %>% 
  arc.select() %>% 
  arc.data2sf() %>% 
  st_transform(crs = 4326)

# URL for the city boundary from the Council's AGOL Open Data
city_boundary_url  <- str_c(
  "https://utility.arcgis.com/usrsvcs/servers/4cdfd020c6f54581a3065c734535adab",
  "/rest/services/AGOL/OpenData/MapServer/11")

# Get the city boundary as a simple feature with WSG84 coordinates 
sf_city_boundary <- arc.open(city_boundary_url) %>% 
  arc.select() %>% 
  arc.data2sf() %>% 
  st_transform(crs = 4326)

# Plot
mapview(sf_secondary_schools, 
        layer.name = "Secondary schools", 
        label = "short_name") %>% 
  addFeatures(sf_city_boundary, weight = 2, 
              color = "black", opacity = 1, fillOpacity = 0)

For the purpose of comparing isochrones created from different network datasets/services and software, we’ll focus on Sheffield Park Academy.

# Get a subset of just Sheffield Park Academy as a simple feature
sf_park <- filter(sf_secondary_schools, school_org == "Sheffield Park Academy")

# Park Academy coordinates
park_coords <- st_coordinates(sf_park)

# Plot
m <- mapview(sf_park, layer.name = "Park Academy", label = "short_name")

setView(m@map, park_coords[1], park_coords[2], zoom = 15)

AGOL logistic service & ArcGIS Pro

I’ve added a feature layer on Portal of isochrones I created using ArcGIS Pro and the AGOL logistic service. There are some very rough notes on what I did in the TakeawaysDataAndApp.docx under “Appendix i) Breadcrumbs - School Isochrones”.

# URL for secondary school isochrones from the Council's Portal (with a token!)
isochrones_url  <- str_c(
  "https://sheffieldcitycouncil.cloud.esriuk.com/server/rest/services/Hosted/",
  "Secondary_school_400m_isochrome/FeatureServer/0?token=SCuzHcThaOEkh6ewOOmp",
  "NYTOa9AckubYSesbQEF9RRM5jM5ruaWfCAkS7hRSJvwBpDtBhJbiCB_eFZH-REj9z1Tu7jOGF2",
  "-atu5AgJGrPMwm8v_VR96qjHySeRHp2hSNGXQ9pAWAu6QCTsOitAIgs_fJD5pZVw4fwxXSg0es",
  "pjtFC0l1tSnExQ2tFyA3UR5zArYLZDV_8DrwMySPs4_0rzwWR4gShptQqPErbwkY078.")

# Get secondary school isochrones as simple features with WSG84 coordinates
sf_isochrones <- arc.open(isochrones_url) %>% 
  arc.select() %>% 
  arc.data2sf() %>% 
  st_transform(crs = 4326)

# Plot
m <- mapview(sf_secondary_schools, 
             layer.name = "Secondary schools", 
             label = "short_name") + 
 mapview(sf_isochrones, layer.name = "Isochrones", label = FALSE)

addFeatures(m, sf_city_boundary, weight = 2, 
            color = "black", opacity = 1, fillOpacity = 0)

Focusing on the isochrone for our example, Sheffield Park Academy:

# Get a subset of just Sheffield Park Academy isochrone as a simple feature
sf_park_isochrone <- filter(sf_isochrones, facilityid == 28)

# Plot
m <- mapview(sf_park, layer.name = "Park Academy", label = "short_name") +
  mapview(sf_park_isochrone, layer.name = "Isochrone")

setView(m@map, park_coords[1], park_coords[2], zoom = 15)

The roundabout on Prince of Wales Road has a walkway that goes underneath it. The AGOL location service hasn’t included this walkway, so the isochrone stops incorrectly short of the Parkway.

openrouteservice & R

The openrouteservice R client package provides easy access to the openrouteservice (ORS) API from R.

# # Get school isochrones from the ORS
# sf_park_ors_isochrone <- ors_isochrones(st_coordinates(sf_park),
#                                    profile = "foot-walking",
#                                    range_type = "distance",
#                                    range = 400,
#                                    output = "sf")

# Point on the road
sf_park_ors_isochrone <- ors_isochrones(c(-1.41661,53.37425),
                                   profile = "foot-walking",
                                   range_type = "distance",
                                   range = 400,
                                   output = "sf")

# Bounding box for Park Academy 
park_bbox <- st_buffer(sf_park_ors_isochrone, 200) %>% 
  st_bbox() %>%
  as.numeric()

# OSM paths
park_paths <- opq(park_bbox) %>%
  add_osm_feature(key = "highway") %>% 
  osmdata_sf() %>% 
  osm_poly2line() 

park_paths <- park_paths$osm_lines %>% 
  filter(is.na(foot) | foot != "no")

# Plot
m <- mapview(sf_park, layer.name = "Park Academy", label = "short_name") +
  mapview(sf_park_ors_isochrone, layer.name = "Isochrone", 
          label = FALSE, alpha.region = 0.2) + 
  mapview(park_paths, layer.name = "Paths")

setView(m@map, park_coords[1], park_coords[2], zoom = 15) %>% 
  addMeasure(primaryLengthUnit = "meters")

Isochrone Notes

Laurie Platt (BI Team, Sheffield City Council)

16 September, 2021

Introduction

Secondary schools

AGOL logistic service & ArcGIS Pro

openrouteservice & R

OSMM Highways Network & Portal

OSMM Highways Network & ArcGIS Pro

Other options