od (version 0.5.1)

od_disaggregate: Split-up each OD pair into multiple OD pairs based on subpoints/subzones

Description

This function is for splitting-up OD pairs. It increases the number of rows in an OD dataset, while aiming to keep the amount of travel represented in the data the same. To take an analogy from another package, it's roughly equivalent to tidyr::pivot_longer().

Usage

od_disaggregate(
  od,
  z,
  subpoints = NULL,
  code_append = "_ag",
  population_column = 3,
  max_per_od = 5,
  keep_ids = TRUE,
  integer_outputs = FALSE
)
od_split(
  od,
  z,
  subpoints = NULL,
  code_append = "_ag",
  population_column = 3,
  max_per_od = 5,
  keep_ids = TRUE,
  integer_outputs = FALSE
)

Arguments

od

An origin-destination data frame

z

Zones representing origins and destinations

subpoints

Points, lines or polygons within the zones. These define the OD data start/end points.

code_append

The name of the column containing aggregate zone names

population_column

The column containing the total population (if it exists)

max_per_od

Maximum flow in the population_column to assign per OD pair. This only comes into effect if there are enough subpoints to choose from.

keep_ids

Should the origin and destination ids be kept? TRUE by default, meaning 2 extra columns are appended, with the names o_agg and d_agg containing IDs from the original OD data.

integer_outputs

Should integer outputs be returned? FALSE by default. Note: there is a known issue when integer results are generated. See https://github.com/ITSLeeds/od/issues/31 for details.

Details

An alias for the function is od_split().

Examples

od = od_data_df[1:2, c(1, 2, 9)]
od
zones = od::od_data_zones_min
od_sf = od_to_sf(od, zones)
set.seed(2021) # for reproducibility
od_disag = od_disaggregate(od, zones)
od_disag2 = od_disaggregate(od, zones, max_per_od = 11)
plot(zones$geometry)
plot(od_sf$geometry, lwd = 9, add = TRUE)
plot(od_disag$geometry, col = "grey", lwd = 1, add = TRUE)
plot(od_disag2$geometry, col = "green", lwd = 1, add = TRUE)
table(od_disag$o_agg, od_disag$d_agg)
# integer results
od_disaggregate(od, zones, integer_outputs = TRUE)

# with more trips per disaggregated OD pair:
disag = od_disaggregate(od_data_df[1:2, ], z = zones, max_per_od = 50)
plot(disag[0])

# with subpoints
subpoints = sf::st_sample(zones, 100)
od_disag_subpoints = od_disaggregate(od, zones, subpoints = subpoints)
plot(subpoints)
plot(od_disag_subpoints$geometry, add = TRUE)

# with buildings data
od_disag_buildings = od_disaggregate(od, zones, od_data_buildings)
summary(od_disag_buildings)
plot(od_data_buildings$geometry)
plot(od_disag_buildings[3], add = TRUE)
# mapview::mapview(od_disag_buildings)

od = od_data_df[1:2, 1:4]
subzones = od_data_zones_small
try(od_disaggregate(od, zones, subzones))
od_disag = od_disaggregate(od, zones, subzones, max_per_od = 500)
ncol(od_disag) - 3 == ncol(od) # same number of columns, the same...
# Except disag data gained geometry and new agg ids:
sum(od_disag[[3]]) == sum(od[[3]])
sum(od_disag[[4]]) == sum(od[[4]])
plot(od_disag)
# test with road network dataset (don't run as time consuming):
if (FALSE) {
od_disag_net = od_disaggregate(od, zones, od_road_network, max_per_od = 500)
plot(zones$geometry)
plot(od_road_network$geometry, add = TRUE, col = "green")
plot(od_disag_net$geometry, add = TRUE)
mapview::mapview(zones) + od_disag_net + od_road_network
}