| Title: | Calculate Travel Times Over Space |
|---|---|
| Description: | Calculate travel time across a landscape from a set of user-specified locations, over a friction surface describing the time taken to cross each cell. Implements the least-cost travel time approach of Weiss et al. (2018) <doi:10.1038/nature25181> and Weiss et al. (2020) <doi:10.1038/s41591-020-1059-1>, and provides convenient access to their global walking and motorised friction surfaces via the 'malariaAtlas' package. Travel times are computed with 'terra'. See Ryan et al. (2025) <doi:10.12688/gatesopenres.16356.1> for further details. |
| Authors: | Gerry Ryan [aut, cre] (ORCID: <https://orcid.org/0000-0003-0183-7630>), Daniel Weiss [aut] (ORCID: <https://orcid.org/0000-0002-6175-5648>), Nicholas Tierney [ctb] (ORCID: <https://orcid.org/0000-0003-1460-8722>) |
| Maintainer: | Gerry Ryan <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 1.0.0 |
| Built: | 2026-10-21 07:02:01 UTC |
| Source: | https://github.com/idem-lab/traveltime |
Calculate the travel time from a set of points over a friction surface.
calculate_travel_time( friction_surface, points, filename = NULL, overwrite = FALSE )calculate_travel_time( friction_surface, points, filename = NULL, overwrite = FALSE )
friction_surface |
A |
points |
A two-column |
filename |
|
overwrite |
|
Implements methods from Weiss et al. 2018, 2020 to calculate travel time from given locations over a friction surface.
Over large areas this function can require significant RAM and will be slow.
Pre-prepared walking or motorised friction surfaces can be obtained
with get_friction_surface(). Users can also provide their own friction
surface. This surface must be in resistance units (1/conductance), e.g.
minutes/meter. See Van Etten 2017.
Note on this implementation: as of package version 1. Earlier versions of
this package used the raster and gdistance packages. These
dependencies have been removed and now instead rely exclusively on terra
for spatial data handling. The two approaches use the same cost model
(eight-direction movement, edge cost distance * mean(friction)) and
produce effectively the same result. terra::costDist() computes true
geodesic inter-cell distances, so the new approach is more accurate than
the old gdistance's distance approximation away from the equator (the
results are identical at the equator and for projected coordinate systems).
The remaining difference is a boundary convention at the origins:
gdistance includes half of the starting cell's own friction when leaving
a point, whereas terra::costDist measures cost to the border of the
origin cells and omits it. This appears as a small, near-constant offset
(about half a cell of friction) close to the origin points. See the
validation article
https://idem-lab.github.io/traveltime/articles/costdist-validation.html
for the full comparison.
Citations:
D. J. Weiss, A. Nelson, C. A. Vargas-Ruiz, K. Gligoric, S., Bavadekar, E. Gabrilovich, A. Bertozzi-Villa, J. Rozier, H. S. Gibson, T., Shekel, C. Kamath, A. Lieber, K. Schulman, Y. Shao, V. Qarkaxhija, A. K. Nandi, S. H. Keddie, S. Rumisha, P. Amratia, R. Arambepola, E. G. Chestnutt, J. J. Millar, T. L. Symons, E. Cameron, K. E. Battle, S. Bhatt, and P. W. Gething. Global maps of travel time to healthcare facilities. (2020) Nature Medicine. doi:10.1038/s41591-020-1059-1
D. J. Weiss, A. Nelson, H.S. Gibson, W. Temperley, S. Peedell, A. Lieber, M. Hancher, E. Poyart, S. Belchior, N. Fullman, B. Mappin, U. Dalrymple, J. Rozier, T.C.D. Lucas, R.E. Howes, L.S. Tusting, S.Y. Kang, E. Cameron, D. Bisanzio, K.E. Battle, S. Bhatt, and P.W. Gething. A global map of travel time to cities to assess inequalities in accessibility in 2015. (2018). Nature. doi:10.1038/nature25181
van Etten, J. (2017). R package gdistance: Distances and routes on geographical grids. Journal of Statistical Software doi:10.18637/jss.v076.i13
A SpatRaster of travel time (in the time units of
friction_surface, e.g. minutes) from each cell to the nearest of
points, with a single layer named "travel_time". Cells that cannot
be reached from any of points (e.g. because they are separated from
them by NA cells in friction_surface) are NA.
# a small synthetic friction surface (minutes per metre), so this example # runs without an internet connection friction_surface <- terra::rast( nrows = 50, ncols = 50, xmin = 111, xmax = 112, ymin = 0, ymax = 1, crs = "EPSG:4326", vals = 0.01 ) from_here <- data.frame( x = c(111.2, 111.9), y = c(0.2, 0.35) ) calculate_travel_time( friction_surface = friction_surface, points = from_here ) # a real motorised friction surface, downloaded from the Malaria Atlas # Project (requires an internet connection) motor_surface <- get_friction_surface( surface = "motor2020", extent = c(111, 112, 0, 1) ) if (!is.null(motor_surface)) { calculate_travel_time( friction_surface = motor_surface, points = from_here ) }# a small synthetic friction surface (minutes per metre), so this example # runs without an internet connection friction_surface <- terra::rast( nrows = 50, ncols = 50, xmin = 111, xmax = 112, ymin = 0, ymax = 1, crs = "EPSG:4326", vals = 0.01 ) from_here <- data.frame( x = c(111.2, 111.9), y = c(0.2, 0.35) ) calculate_travel_time( friction_surface = friction_surface, points = from_here ) # a real motorised friction surface, downloaded from the Malaria Atlas # Project (requires an internet connection) motor_surface <- get_friction_surface( surface = "motor2020", extent = c(111, 112, 0, 1) ) if (!is.null(motor_surface)) { calculate_travel_time( friction_surface = motor_surface, points = from_here ) }
SpatRaster or SpatVector
Formats spatial extent for use in get_friction_surface.
ext_from_terra(r)ext_from_terra(r)
r |
|
2x2 matrix
library(terra) r <- terra::rast( extent = terra::ext(c(111, 112, 0, 1)) ) ext_from_terra(r)library(terra) r <- terra::rast( extent = terra::ext(c(111, 112, 0, 1)) ) ext_from_terra(r)
This function allows get_friction_surface() to accept the argument extent as a vector, 2x2 matrix, SpatVector, or SpatRaster and extracts and converts into the annoyingly specific format necessary to download the friction surface. See details in ?get_friction_surface
ext_matrix(extent)ext_matrix(extent)
extent |
|
2x2 matrix with column names "x" and "y" and row names "min"
and "max"
# vector/double x <- c(111, 112, 0, 1) ext_matrix(x) # SpatExtent y <- terra::ext(x) ext_matrix(y) # SpatRaster r <- terra::rast(extent = y) ext_matrix(r) # SpatVector v <- terra::vect(y) ext_matrix(v)# vector/double x <- c(111, 112, 0, 1) ext_matrix(x) # SpatExtent y <- terra::ext(x) ext_matrix(y) # SpatRaster r <- terra::rast(extent = y) ext_matrix(r) # SpatVector v <- terra::vect(y) ext_matrix(v)
Extent vector to matrix
ext_vect_to_matrix(x)ext_vect_to_matrix(x)
x |
|
2x2 matrix
ext_vect_to_matrix(c(111,112,0, 1))ext_vect_to_matrix(c(111,112,0, 1))
Wrapper function to download friction surfaces via
malariaAtlas::getRaster().
get_friction_surface( surface = c("motor2020", "walk2020"), filename = NULL, overwrite = FALSE, extent = NULL )get_friction_surface( surface = c("motor2020", "walk2020"), filename = NULL, overwrite = FALSE, extent = NULL )
surface |
|
filename |
|
overwrite |
|
extent |
Spatial extent as one of:
See details |
Convenience wrapper to malariaAtlas::getRaster() to access
motorised and walking travel friction layers per Weiss et al. 2020, that
adds safety to check existing files before download. Surfaces can be
downloaded directly from:
https://malariaatlas.org/project-resources/accessibility-to-healthcare/.
Here —
surface = "motor2020" will download
"Accessibility__202001_Global_Motorized_Friction_Surface", and
surface = "walk2020" will download
"Accessibility__202001_Global_Walking_Only_Friction_Surface".
D. J. Weiss, A. Nelson, C. A. Vargas-Ruiz, K. Gligoric, S., Bavadekar, E. Gabrilovich, A. Bertozzi-Villa, J. Rozier, H. S. Gibson, T., Shekel, C. Kamath, A. Lieber, K. Schulman, Y. Shao, V. Qarkaxhija, A. K. Nandi, S. H. Keddie, S. Rumisha, P. Amratia, R. Arambepola, E. G. Chestnutt, J. J. Millar, T. L. Symons, E. Cameron, K. E. Battle, S. Bhatt, and P. W. Gething. Global maps of travel time to healthcare facilities. (2020) Nature Medicine. doi:10.1038/s41591-020-1059-1
extent is formatted and then passed through to
malariaAtlas::getRaster() as a 2x2 matrix. It is converted into a matrix
using ext_matrix. The matrix format used is as returned by
sf::st_bbox() — the first column has the minimum, the second the
maximum values; rows 1 & 2 represent the x & y dimensions respectively:
matrix(c("xmin", "ymin", "xmax", "ymax"), nrow = 2, ncol = 2, dimnames = list(c("x", "y"), c("min", "max"))). extent must be supplied.
Troubleshooting: this function requires an internet connection. If the
download fails (e.g. a warning Failed to connect to MAP geoserver), a
warning is given and NULL is returned. This is an issue with fetching
the friction surface using malariaAtlas::getRaster(), and not an issue
with traveltime itself. The server may be down or there may be an issue
with that package.
A SpatRaster friction surface with a single layer named
"friction_surface", or NULL (with a warning) if the download fails.
# requires an internet connection # for more examples of passing extent types see ?ext_matrix ext <- c(111, 112, 0, 1) get_friction_surface( surface = "motor2020", extent = ext )# requires an internet connection # for more examples of passing extent types see ?ext_matrix ext <- c(111, 112, 0, 1) get_friction_surface( surface = "motor2020", extent = ext )
Locations of all MRT and LRT station exits in Singapore.
stationsstations
A matrix with 563 rows and 2 columns:
longitude
latitude
Contains information from the dataset "LTA MRT Station Exit (GEOJSON)" accessed on the 10th of December 2024 from data.gov.sg (https://data.gov.sg/datasets/d_b39d3a0871985372d7e1637193335da5/view), which is made available under the terms of the Singapore Open Data Licence version 1.0 https://data.gov.sg/open-data-licence.
Land Transport Authority. (2019). LTA MRT Station Exit (GEOJSON) (2024) Dataset. data.gov.sg. Retrieved December 10, 2024 from https://data.gov.sg/datasets/d_b39d3a0871985372d7e1637193335da5/view