The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.

movecost 3.0

Calculation of slope-dependent accumulated cost surfaces, least-cost paths, least-cost corridors, least-cost networks, ranked alternative routes, cost allocation, and cost boundaries related to human movement across the landscape, with 26 selectable cost functions. See Alberti (2019) doi:10.1016/j.softx.2019.100331 for an earlier version of the package.

Version 3.0 is a ground-up architectural redesign of the 2.x series. The analytical capabilities and the cost functions are retained; what changes is how the package computes, scales, and visualises. See NEWS.md for the full list of changes and the 2.x-to-3.0 function correspondence table.

Design in one paragraph

Build the cost graph once with mc_surface(); reuse it with every analysis function; visualise on demand with plot(). Computation and visualisation are fully decoupled: every function returns an object that stores its results, and the plot()/autoplot() methods (ggplot2) can be invoked, customised, and re-invoked at any time without re-running anything.

Quick start

library(movecost)

# sample data shipped with the package
dtm    <- mc_volc()          # SpatRaster
start  <- mc_volc_loc()      # sf points
destin <- mc_destin_loc()    # sf points

# 1. build the cost surface ONCE (Tobler's on-path hiking function)
surf <- mc_surface(dtm, funct = "t", move = 16)

# 2. reuse it across analyses - nothing is recomputed
acc <- mc_accum(surf, origin = start, breaks = 0.05)   # isochrones
lcp <- mc_paths(surf, origin = start, destin = destin) # least-cost paths
nw  <- mc_network(surf, nodes = destin)                # LCP network + cost matrix
al  <- mc_alloc(surf, origin = destin)                 # cost allocation
rk  <- mc_rank(surf, start, destin[2, ], k = 4)        # ranked alternatives
bd  <- mc_boundary(surf, destin[1:3, ], limit = 5, time = "m")
co  <- mc_corridor(surf, a = destin[1, ], b = destin[4, ])

# 3. visualise on demand; plots are ggplot objects, customise at will
plot(acc)
p <- plot(lcp) + ggplot2::ggtitle("My custom title")
p
ggplot2::ggsave("lcp.pdf", p, width = 7, height = 7)

# compare cost functions
cmp <- mc_comp(dtm, start, destin[2, ], functs = c("t", "ks", "hrz"))
plot(cmp); plot(cmp, type = "chart")

# the 26 implemented cost functions, annotated
mc_cost_functions()

# persistence and GIS export
mc_save(surf, "surf.rds"); surf <- mc_load("surf.rds")
mc_export(lcp, dir = "outputs")

Installation (from this source folder)

install.packages(c("terra", "sf", "igraph", "ggplot2", "devtools"))
devtools::install()        # run from the package root
# optional extras:
install.packages("elevatr")   # online DTM download via mc_dtm()

To (re)generate the help files from the roxygen2 annotations and run the full quality pipeline:

devtools::document()   # builds man/*.Rd from the roxygen2 annotations
devtools::test()       # runs the analytic + consistency test suite
devtools::check()      # full R CMD check

Why 3.0 is faster and more scalable

  1. Compute-once architecture: the conductance graph is built a single time per cost function; in 2.x it was rebuilt inside every call (an n-location network implied O(n) redundant constructions).
  2. Batched multi-source queries: allocation, boundaries, and cost matrices are served by single igraph Dijkstra calls over all origins.
  3. Modern stack: terra (C++-backed raster I/O and algebra) replaces raster/sp; igraph replaces the gdistance accumulation machinery; chron is replaced by an internal formatter. Legacy sp/raster inputs are still accepted and converted automatically.
  4. NoData-aware graph: cells with NoData never enter the graph, so irregular DTMs (coastlines) need no special handling - and the graph gets smaller, not larger, when the study area is irregular.

Validation

The computational core was validated against analytic results on synthetic terrains, against internal consistency laws, and against the 2.x (gdistance-based) procedure: see inst/validation/ and the testthat suite. Two deliberate methodological corrections (corridor directionality; signed-slope handling in the ks and ma cost functions) are documented in NEWS.md and in ?mc_cost_functions.

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.