Scottish Cycling by Design classification • osmactive

options(timeout = 3000)
library(osmactive)
library(tidyverse)
library(sf)
tmap::tmap_mode("view")

The default classification is the Cycling by Design guidance from the Scottish Government. This is a detailed classification system that includes the following categories, based on a minimal example dataset:

osm = osm_edinburgh
cycle_net = get_cycling_network(osm)
drive_net = get_driving_network(osm)
drive_net_major = get_driving_network(osm)
cycle_net = distance_to_road(cycle_net, drive_net)
cycle_net = classify_cycle_infrastructure(cycle_net)
table(cycle_net$detailed_segregation)
table(cycle_net$cycle_segregation)

m = plot_osm_tmap(cycle_net)
m

Let’s try running the code on a larger example, covering Edinburgh:

osm = get_travel_network("Edinburgh")
cycle_net = get_cycling_network(osm)
drive_net = get_driving_network(osm)
cycle_net = distance_to_road(cycle_net, drive_net)
cycle_net = classify_cycle_infrastructure(cycle_net)
central_edinburgh = osm_edinburgh |>
  sf::st_union() |>
  sf::st_centroid() |>
  sf::st_buffer(3000)
m = plot_osm_tmap(cycle_net[central_edinburgh, ])
m

We can pick out specific bits of infrastructure, e.g. those mentioned in issue #67:

https://www.openstreetmap.org/way/1137759929 shows up as this right now, but this looks like a narrow segregated / stepped track to me. https://www.openstreetmap.org/way/1209807768 is another case that doesn’t seem like off-road.

My idea of off-road cycleway were things like https://www.openstreetmap.org/way/7973833 and https://www.openstreetmap.org/way/265320503 – things with no parallel road nearby.

Let’s pull out those bits of infrastructure:

cycle_net_included = cycle_net |>
  subset(osm_id %in% c(1137759929, 1209807768, 7973833, 265320503))
names(cycle_net_included)
cycle_net_included[c("osm_id", "distance_to_road", "cycle_segregation")] |>
  sf::st_drop_geometry() |>
  knitr::kable()

We can also plot these as follows:

m = plot_osm_tmap(cycle_net_included)
m

Level of service

The level of service associated with each link can be calculated based on the Level of Service table in the Cycling by Design guidance:

We can then create the Speed Limit (mph) and AADT columns in the cycle_net object:

cycle_net_speeds = clean_speeds(cycle_net)
cycle_net_volumes = estimate_traffic(cycle_net_speeds)
cycle_net$`Speed Limit (mph)` = classify_speeds(cycle_net_speeds$maxspeed_clean)
cycle_net$AADT = npt_to_cbd_aadt_numeric(cycle_net_volumes$assumed_volume)
table(cycle_net$`Speed Limit (mph)`)
table(cycle_net$AADT)

Finally, we can calculate the level of service for each link:

# devtools::load_all()
cycle_net_los = level_of_service(cycle_net) 
# |>
# select(osm_id, `Level of Service`, `Speed Limit (mph)`, AADT, cycle_segregation)
table(cycle_net_los$`Level of Service`)
# Count NA values:
cycle_net_los |>
  sf::st_drop_geometry() |>
  group_by(`Speed Limit (mph)`, AADT, cycle_segregation) |>
  summarise(na = sum(is.na(`Level of Service`))) |>
  filter(na > 0) |>
  arrange(desc(na))
cycle_net_los_plot = cycle_net_los |>
  select(osm_id, `Speed Limit (mph)`, AADT, cycle_segregation, `Level of Service`)
# mapview::mapview(cycle_net_los_plot, zcol = "Level of Service")