rudeboybert / spatialepi

R Package of methods and data for spatial epidemiology

SpatialEpi

Package of data and methods for spatial epidemiology.

Installation

Get the released version from CRAN:

install.packages("SpatialEpi")

Or the development version from GitHub:

# If you haven't installed devtools yet, do so:
# install.packages("devtools")
devtools::install_github("rudeboybert/SpatialEpi")

Note: In order for all C++ code to compile correctly you may need to

1. Install the cpp11 package

2. Install an older version of RcppArmadillo by running

   packageurl <- "https://cran.r-project.org/src/contrib/Archive/RcppArmadillo/RcppArmadillo_0.9.900.3.0.tar.gz"
   install.packages(packageurl, repos=NULL, type="source")

Example

We load the data and convert the coordinate system from latitude/longitude to a grid-based system.

library(SpatialEpi)

data(NYleukemia)
sp.obj <- NYleukemia$spatial.polygon
centroids <- latlong2grid(NYleukemia$geo[, 2:3])
population <- NYleukemia$data$population
cases <- NYleukemia$data$cases

We plot the incidence of leukemia for each census tract.

plotmap(cases/population, sp.obj, log=TRUE, nclr=5)
points(grid2latlong(centroids), pch=4)

We run the Bayesian Cluster Detection method from Wakefield and Kim (2013):

y <- cases
E <- expected(population, cases, 1)
max.prop <- 0.15
shape <- c(2976.3, 2.31)
rate <- c(2977.3, 1.31)
J <- 7
pi0 <- 0.95
n.sim.lambda <- 10^4
n.sim.prior <- 10^5
n.sim.post <- 10^5

# Compute output
output <- bayes_cluster(y, E, population, sp.obj, centroids, max.prop,
                        shape, rate, J, pi0, n.sim.lambda, n.sim.prior,
                        n.sim.post)
#> [1] "Algorithm started on: Thu Jun 10 16:38:10 2021"
#> [1] "Importance sampling of lambda complete on: Thu Jun 10 16:39:11 2021"
#> [1] "Prior map MCMC complete on: Thu Jun 10 16:43:03 2021"
#> [1] "Posterior estimation complete on: Thu Jun 10 16:56:47 2021"

plotmap(output$post.map$high.area, sp.obj)
#> Warning in wkt(obj): CRS object has no comment

#> Warning in wkt(obj): CRS object has no comment

spatialepi's Issues

retirement of rgdal, rgeos and maptools

as specified here

Fix rccp error locally

The expected calculation method in the SpatialEpi package

Hi, I am curious about the formula for the expected number of incidences in the SpatialEpi package.
I would be honored if I could get your reply.

Transfer Data Examples from old vignette to new manuscript

Do demos of

• pennLC
• NYleukemia
• scotland

using new sf code

Update install_geo_libraries vignette

• You really must not recommend the use of MacPorts. Proper, static
  builds of gdal and geos are available from mac.r-project.org, and binary
  macOS packages for rgeos and rgdal are available, And you do not need
  to use a terminal to download and install packages: it can and probably
  should be done from inside R, either via install.packages() or from the
  GUI menus.

Similarly, Windows users can just use binary packages. OSGeo4W is not
needed even when installing from source.

Then vignette back into package

Setup skeleton of manuscript

• Make sure peerJ rticles template works
• Change basic author info

New GDAL/PROJ require UTM zone

In revdeps SpatialEpi fails in mapvariable() in ?EBpostthresh:
00check.log
because "+proj=utm" has never been valid without a zone, and now a zone number is required. See also GlobalEcologyLab/poems#1, PredictiveEcology/NetLogoR#43:

> slot(scotland$spatial.polygon, "proj4string")
Coordinate Reference System:
Deprecated Proj.4 representation: +proj=utm 
Warning message:
In wkt(x) : CRS object has no comment

Fix Error: package or namespace load failed for 'SpatialEpi' in dyn.load(file, DLLpath = DLLpath, ...)

unable to load shared object '/Library/Frameworks/R.framework/Versions/4.0/Resources/library/00LOCK-SpatialEpi/00new/SpatialEpi/libs/SpatialEpi.so':
dlopen(/Library/Frameworks/R.framework/Versions/4.0/Resources/library/00LOCK-SpatialEpi/00new/SpatialEpi/libs/SpatialEpi.so, 6): Library not loaded: /Library/Frameworks/R.framework/Versions/3.5/Resources/lib/libR.dylib
Referenced from: /Library/Frameworks/R.framework/Versions/4.0/Resources/library/00LOCK-SpatialEpi/00new/SpatialEpi/libs/SpatialEpi.so
Reason: image not found
Error: loading failed
Execution halted
ERROR: loading failed

• removing '/Library/Frameworks/R.framework/Versions/4.0/Resources/library/SpatialEpi'
• restoring previous '/Library/Frameworks/R.framework/Versions/4.0/Resources/library/SpatialEpi'

Error in kulldorff function for cincinnati data

Hello, I'm working on the Cincinnati data from the paper "Grubesic, T, R. Wei and A. Murray (2014). Spatial clustering overview and comparison: accuracy, sensitivity and computational expense. Annals, Association of American Geographers 104, 1134-1156.", available at GeoDat github repo.

This is the format of the data:

> str(cincinnati, max.level = 1)
    Classes ‘sf’ and 'data.frame':	457 obs. of  73 variables:
    $ ID        : num  726907 695744 695762 695780 695798 ...
    $ POPULATION: num  479 85 29 117 96 45 81 32 41 89 ...
    $ ASSAULT   : int  0 2 0 3 0 9 1 1 0 1 ...
    $ geometry  :sfc_POLYGON of length 457; first list element: List of 1

I assigned the arguments geo and expected.cases like this:

geo = cincinnati %>% group_by(ID) %>% summarise(st_centroid(geometry)) %>% st_coordinates() %>% as.data.frame()
EC = expected(cincinnati$POPULATION, cincinnati$ASSAULT, 1) # I used  n.strata = 1 because it returns a vector of 457 elements (just like cincinnati$ASSAULT or cincinnati$POPULATION)

Now, when I call the function kulldorff, i get the error: "Error in hist.default(combined.lambda, main = "Monte Carlo Distribution of Lambda", : invalid number of 'breaks' "

poisson <- kulldorff(geo,                     # an n x 2 table of the (x,y)-coordinates of the area centroids
                     cincinnati$ASSAULT,      # aggregated case counts for all n areas
                     cincinnati$POPULATION,   # aggregated population counts for all n areas
                     expected.cases=EC,       # expected numbers of disease for all n areas 
                     pop.upper.bound=0.5,     # the upper bound on the proportion of the total population each zone can include
                     n.simulations=999,       # number of Monte Carlo samples used for significance measures
                     alpha.level=0.05,        # alpha-level threshold used to declare significance
                     plot=TRUE                # flag for whether to plot histogram of Monte Carlo samples of the LL of the most likely cluster
                     )

If I change the argument plot = FALSE, i get the error: "Error in if (new.secondary.cluster$p.value > alpha.level) { : missing value where TRUE/FALSE needed "

At this point I don't understand how can I fix it from the arguments, can you please help me indicating how can I solve it?, finally, my setup enviroment in R is:

R version 4.2.3 (2023-03-15)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 20.04.5 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/liblapack.so.3

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] tools     stats     graphics  grDevices utils     datasets  methods  
[8] base     

other attached packages:
[1] SpatialEpi_1.2.8 sp_1.6-0         sf_1.0-12        stringr_1.4.1   
[5] ggplot2_3.4.1    dplyr_1.1.1     

loaded via a namespace (and not attached):
 [1] spdep_1.2-8        Rcpp_1.0.10        pillar_1.9.0       compiler_4.2.3    
 [5] class_7.3-21       boot_1.3-28        lattice_0.20-45    lifecycle_1.0.3   
 [9] tibble_3.2.1       gtable_0.3.3       pkgconfig_2.0.3    rlang_1.1.0       
[13] DBI_1.1.3          cli_3.6.1          spData_2.2.2       e1071_1.7-13      
[17] s2_1.1.2           withr_2.5.0        generics_0.1.3     vctrs_0.6.1       
[21] classInt_0.4-9     grid_4.2.3         tidyselect_1.2.0   glue_1.6.2        
[25] R6_2.5.1           fansi_1.0.4        deldir_1.0-6       farver_2.1.1      
[29] magrittr_2.0.3     MASS_7.3-58.3      scales_1.2.1       units_0.8-1       
[33] colorspace_2.1-0   labeling_0.4.2     utf8_1.2.3         KernSmooth_2.23-20
[37] stringi_1.7.12     proxy_0.4-27       wk_0.7.2           munsell_0.5.0    

Add testthat unit tests

Add testthat unit tests on on kulldorff, besag, bayesian methods on a simple/toy example for README

Update instructions for installing `gdal` and `geos` geospatial libraries if necessary

in vignettes/install_geo_libraries.Rmd

Convert mapvariable() to plot sf objects

Refactor expected() function

In a new branch off SURF21:

• Write test_that unit tests: https://r-pkgs.org/tests.html. Example
• Create expected_new() function, assume tidy data frame input for population and cases, convert code to tidyverse
• Copy over documentation and format in roxygen2: https://r-pkgs.org/man.html
• Be sure to install & check package early and often

Create PR early, then tag me for review when done

Create rticles skeleton for PeerJ

Manuscript Error

Quitting from lines 139-144 (Manuscript.Rmd) Error in if (plot) { : argument is not interpretable as logical Calls: <Anonymous> ... withCallingHandlers -> withVisible -> eval -> eval -> kulldorff Execution halted

Write about pennLC in manuscript:

• Talk about data source, year, geographic level, etc. i.e. give data's context (Use Pennsylvania Department of Health and link to new website for context)
• Show users R code of how they could make static choropleth map of lung cancer cases

Update authors in DESCRIPTION file

add myself as an author

Convert all sp spatial data to sf package objects

• pennLC. Arris will wait for Bert review.
• Scotland
• NY leukemia

sp is the old version of sf package for maps from 192. Currently pennLC data PA lung cancer data is saved very disorganized as follows:

library(SpatialEpi)
library(sp)
# case and population data
pennLC$data
# map data
pennLC$spatial.polygon

Goal: Create a sf data frame that allows you to create a choropleth map of the above data.

1. Download a new shapefile of Pennsylvania that is county level
2. Or convert PennLC$spatial.polygon into SF object
3. Ensure that map and data-frame is attached (PennLC is the map)

As reference use https://r-pkgs.org/data.html?q=data-raw#data-sysdata
As example of workflow for adding data to R package, use 538:

• You will only add data via data-raw/ folder
• In particular this example: https://github.com/rudeboybert/fivethirtyeight/blob/master/data-raw/process_data_sets.R#L12
• In your case, in data-raw/ folder, do all manipulations to create the above sf object
• use usethis::use_data() to write to R package

Deliverable is an sf object that user can easily use to plot lung cancer data. target wednesday

Priority: Convert existing `man/*.Rd` old school help files into new school help files using `roxygen2`

• Create a new branch and pull request
• Convert ONE function to the updated roxygen2 format
• Tag Albert when done so he can review
• Wait for @rudeboybert to fix issue