OD data disaggregation

remotes::install_dev("tmap")
library(spanishoddata)
library(tidyverse)
library(sf)
library(tmap)

1 Introduction

This vignette demonstrates origin-destination (OD) data disaggregation using the {odjitter} package. The package is an implementation of the method described in the paper “Jittering: A Computationally Efficient Method for Generating Realistic Route Networks from Origin-Destination Data” (Lovelace, Félix, and Carlino 2022) for adding value to OD data by disaggregating desire lines. This can be especially useful for transport planning purposes in which high levels of geographic resolution are required (see also the od2net for direct network generation from OD data).

2 Data preparation

We’ll start by loading a week’s worth of origin-destination data for the city of Salamanca, building on the example in the README (note: these chunks are not evaluated):

od_db <- spod_get(
  type = "od",
  zones = "distritos",
  dates = c(start = "2024-03-01", end = "2024-03-07")
)
distritos <- spod_get_zones("distritos", ver = 2)
distritos_wgs84 <- distritos |>
  sf::st_simplify(dTolerance = 200) |>
  sf::st_transform(4326)
od_national_aggregated <- od_db |>
  group_by(id_origin, id_destination) |>
  summarise(Trips = sum(n_trips), .groups = "drop") |>
  filter(Trips > 500) |>
  collect() |>
  arrange(desc(Trips))
od_national_aggregated
od_national_interzonal <- od_national_aggregated |>
  filter(id_origin != id_destination)
salamanca_zones <- zonebuilder::zb_zone("Salamanca")
distritos_salamanca <- distritos_wgs84[salamanca_zones, ]
ids_salamanca <- distritos_salamanca$id
od_salamanca <- od_national_interzonal |>
  filter(id_origin %in% ids_salamanca) |>
  filter(id_destination %in% ids_salamanca) |>
  arrange(Trips)
od_salamanca_sf <- od::od_to_sf(
  od_salamanca,
  z = distritos_salamanca
)

3 Disaggregating desire lines

For this you’ll need some additional dependencies:

remotes::install_github("dabreegster/odjitter", subdir = "r")
remotes::install_github("nptscot/osmactive")

We’ll get the road network from OSM:

salamanca_boundary <- sf::st_union(distritos_salamanca)
osm_full <- osmactive::get_travel_network(salamanca_boundary)

osm <- osm_full[salamanca_boundary, ]
drive_net <- osmactive::get_driving_network(osm)
drive_net_major <- osmactive::get_driving_network_major(osm)
cycle_net <- osmactive::get_cycling_network(osm)
cycle_net <- osmactive::distance_to_road(cycle_net, drive_net_major)
cycle_net <- osmactive::classify_cycle_infrastructure(cycle_net)
map_net <- osmactive::plot_osm_tmap(cycle_net)
map_net

We can use the road network to disaggregate the desire lines:

od_jittered <- odjitter::jitter(
  od_salamanca_sf,
  zones = distritos_salamanca,
  subpoints = drive_net,
  disaggregation_threshold = 1000,
  disaggregation_key = "Trips"
)

Let’s plot the disaggregated desire lines:

od_jittered |>
  arrange(Trips) |>
  ggplot() +
  geom_sf(aes(colour = Trips), size = 1) +
  scale_colour_viridis_c() +
  geom_sf(data = drive_net_major, colour = "black") +
  theme_void()

The results show that you can add value to the OD data by disaggregating the desire lines with the {odjitter} package. This can be useful for understanding the spatial distribution of trips within a zone for transport planning.

We have plotted the disaggregated desire lines on top of the major road network in Salamanca. A next step could be routing to help prioritise infrastructure improvements.

Lovelace, Robin, Rosa Félix, and Dustin Carlino. 2022. “Jittering: A Computationally Efficient Method for Generating Realistic Route Networks from Origin-Destination Data.” Findings.