This aims at providing a set of fast, memory efficient functions to perform spatial interaction modelling, also called gravity modelling. Currently, the doubly and singly constrained models are implemented for canonical set of constraints. Future versions will aim to implement more origin and destination constraints as well. It was developed in the context of studying commuter flows by active travel (cycling & walking ) in Great Britain as part of a project at CASA, UCL.
Not yet on CRAN, so please install the development version of
cppSim
with:
# install.packages(C("devtools","pak"))
::install_github("ischlo/cppSim")
devtools# pak::pak("ischlo/cppSim")
The package comes with sample data sets that allow to test the functions right away as well as see the type of input that is recommended.
cppRouting
package.Refer to the vignette to find some theory on SIMs and a
naive implementation in R
.
Using the built-in data sets flows_test
and
distance_test
, we can run a test by following the example
This is a basic example which shows you how to solve a common
problem:
library(cppSim)
## basic example code
data("flows_test")
data("distance_test")
<- run_model(
model_test flows = flows_test,
distance = distance_test
)
For an example of what can be done with this package, please refer to the publication on active travel spatial interaction models in London for which it was originally developed.
The accompanying code for the analysis is provided in the ischlo/quant_cycle_walk
repository.
This package has some dependencies that might need manual installation, although the most important external ones have been provided with the source code.
The package uses the armadillo
library,
which is imported and linked automatically when the package is
installed.
On the R side, it uses Rcpp
[@eddelbuettel2011] and
RcppArmadillo
[@eddelbuettel2014].
Compared to the equivalent functions implemented in pure R, it runs about x10 faster for a \(`\sim 1000 \times 1000`\) OD matrix, the speed up is increasingly more significant as matrices get bigger.
[1] "/Users/cenv1069/Library/R/arm64/4.4/library/cppSim/extdata/benchmark_test.rds"
To cite package 'cppSim' in publications use:
Schlosser I (2023). _cppSim: Fast and Memory Efficient Spatial
Interaction Models_. doi:10.5281/zenodo.14367756
<https://doi.org/10.5281/zenodo.14367756>,
<https://ischlo.github.io/cppSim/>.
A BibTeX entry for LaTeX users is
@Manual{,
title = {cppSim: Fast and Memory Efficient Spatial Interaction Models},
author = {Ivann Schlosser},
year = {2023},
doi = {10.5281/zenodo.14367756},
url = {https://ischlo.github.io/cppSim/},
}