--- title: "OpenRange R package" author: "Brian Maitner" date: "`r Sys.Date()`" output: rmarkdown::html_vignette vignette: > %\VignetteIndexEntry{OpenRange R package} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8} --- # Open Range Open Range is a database containing Open Access species range maps and associated metadata. The database is currently in a pre-alpha, proof-of-concept stage. While the methods underlying the range maps are documented in the metadata, and the maps follow reasonable best practices, they have not been peer-reviewed. They are provided to show a working example of what could be done with an Open Access range database, and to provide an example of what the underlying database might look like. The maps should not be used without critical evaluation. ## Licensing The maps currently contained in the Open Range database are available with a CC-BY license, but we note that additional maps using alternative licenses may be applied in the future. ### Accessing the maps The easiest way to access the species range maps is by querying a species name. OpenRange relies on an internet connection to a remote database. If that database is unavailable, OpenRange functions return `NULL` along with an informative message rather than throwing an error, so it is worth checking the output before using it. ```{r load species} library(OpenRange) pcontorta <- OpenRange_load_species(species = "Pinus contorta") if(!is.null(pcontorta)){ plot(pcontorta[2]) } ``` ```{r availability, include = FALSE} # Skip the remaining chunks if the database could not be reached (or if the # packages used below are not installed) so that the vignette can still be built # when the internet resource is unavailable. openrange_available <- !is.null(pcontorta) && requireNamespace("tidyverse", quietly = TRUE) knitr::opts_chunk$set(eval = openrange_available) ``` ```{r availability note, eval = TRUE, echo = FALSE, results = "asis"} if(!openrange_available){ cat("**Note:** the OpenRange database was not available when this vignette", "was built, so the code below is shown but was not run.\n") } ``` ### Checking map quality This is a bit cumbersome, but keep in mind this is a pre-alpha version. For now, you can use the function OpenRange_get_stats ```{r get metadata} library(tidyverse) # Grab the stats model_stats <- OpenRange_get_stats() if(!is.null(model_stats)){ # Reformat them into a single table model_stats_wide <- bind_rows(model_stats$rangebagging %>% mutate(algorithm = "rangebagging"), model_stats$ppm %>% mutate(algorithm = "ppm"), model_stats$points %>% mutate(algorithm = "points")) # Get relevant model stats model_stats_wide %>% filter(model_id == pcontorta$model_id)->test print(test %>% select(cv_mean_train_npresence,cv_mean_test_n_presence,cv_mean_test_auc,cv_mean_train_auc)) } ``` .