Calculate areas within network distance

[Robinlovelace]

This is a request I received to solve a common problem: identify all areas within some 'network distance' of points.

[Robinlovelace]

Starting point in reprex below:

# Aim: calculate areas within network distance

unzip("~/repos/stplanr/Network_Hassall.zip")
list.files("Network_Hassall/")
#>  [1] "accesspoints.cpg"                             
#>  [2] "accesspoints.dbf"                             
#>  [3] "accesspoints.prj"                             
#>  [4] "accesspoints.sbn"                             
#>  [5] "accesspoints.sbx"                             
#>  [6] "accesspoints.shp"                             
#>  [7] "accesspoints.shp.MAN-NTW-38.6084.4152.sr.lock"
#>  [8] "accesspoints.shp.xml"                         
#>  [9] "accesspoints.shx"                             
#> [10] "BowlingPark3.CPG"                             
#> [11] "BowlingPark3.dbf"                             
#> [12] "BowlingPark3.prj"                             
#> [13] "BowlingPark3.sbn"                             
#> [14] "BowlingPark3.sbx"                             
#> [15] "BowlingPark3.shp"                             
#> [16] "BowlingPark3.shp.MAN-NTW-38.6084.4152.sr.lock"
#> [17] "BowlingPark3.shp.xml"                         
#> [18] "BowlingPark3.shx"                             
#> [19] "SE_Road.dbf"                                  
#> [20] "SE_Road.prj"                                  
#> [21] "SE_Road.sbn"                                  
#> [22] "SE_Road.sbx"                                  
#> [23] "SE_Road.shp"                                  
#> [24] "SE_Road.shp.ed.lock"                          
#> [25] "SE_Road.shp.MAN-NTW-38.6084.4152.sr.lock"     
#> [26] "SE_Road.shx"
roads = sf::read_sf("Network_Hassall/SE_Road.shp")
apnts = sf::read_sf("Network_Hassall/accesspoints.shp")
plot(apnts)

apnt1 = apnts[1, ]

sf::st_crs(apnts)
#> Coordinate Reference System:
#>   EPSG: 27700 
#>   proj4string: "+proj=tmerc +lat_0=49 +lon_0=-2 +k=0.9996012717 +x_0=400000 +y_0=-100000 +ellps=airy +towgs84=446.448,-125.157,542.06,0.15,0.247,0.842,-20.489 +units=m +no_defs"
sf::st_crs(roads)
#> Coordinate Reference System:
#>   EPSG: 27700 
#>   proj4string: "+proj=tmerc +lat_0=49 +lon_0=-2 +k=0.9996012717 +x_0=400000 +y_0=-100000 +ellps=airy +towgs84=446.448,-125.157,542.06,0.15,0.247,0.842,-20.489 +units=m +no_defs"

a_buff = sf::st_buffer(apnt1, 1000)
r_buff = sf::st_intersection(roads, a_buff)
#> Warning: attribute variables are assumed to be spatially constant
#> throughout all geometries

mapview::mapview(r_buff) + mapview::mapview(apnt1)

^{Created on 2019-06-12 by the reprex package (v0.3.0)}

[mpadge]

Depends what you mean by "areas"? The new dodgr function, dodgr_fundamental_cycles enables all areas enclosed by cycles of highways to be identified. Or do you mean something else?

[Robinlovelace]

Hi @mpadge thanks for spotting this and the quick response. I was thinking that dodgr could be a suitable (and fast!) way to solve this. Thinking about it, I believe it would be sufficient to return the network geometries associated with each point, i.e. something like this:

[Robinlovelace]

Then a buffer operation, with assumptions about feasible distance to walk to roads, could be used to convert the result into a polygon that can be overlaid on a residential zoning system to estimate the number of people within x km network distance to each point (in this case parks). Would dodgr_fundamental_cycles() be an appropriate tool for the job? The network, provided by @Katatrepsis, is quite large and there are many points.

I plan to upload the input data as it's all open access I believe. Good benchmark!

[Robinlovelace]

Also hearing from @mem48 that the in-development https://github.com/ITSLeeds/opentripplanner has good any modal isochrones.

[mem48]

It might be better to get all nodes within a ditance with dodgr and do a concave hull on the points