Creating country checklists using GBIF and rgbif

I wrote this function in order to create simple summaries of GBIF data per country (i.e. checklists based on the available GBIF information) to supply to other people without R expertise. I suspect there is a lot that could be improved (and perhaps I have even missed something obvious within the rgbif package that would have made it easier). Anyhow, this is where it is at. I have been using the function 'lapply-ed' across a vector of country ISO 2 character codes. Occasionally it fails (particularly with countries where there is a lot of data), hence the try...catch approach for when applying over such lists.

Of course, I realise that GBIF is not comprehensive in its coverage, and that there are often issues that need addressing with occurrence data and taxonomic precision; this code was written for a quick first assessment useful for comparing the presence/absence of potentially invasive species in countries in the context of horizon scanning. Use this code at your own risk!

Feedback/improvements welcome!

	#### Function for retrieving GBIF country-based checklists based on existing records
	# 01.04.2018, OL Pescott
	## v0, Script for retrieving lists of taxa per country from GBIF (not as straightforward as you might think!), O.Pescott, 22.03.2018
	#rm(list = ls())
	# Potential improvements:
	# Allow for more arguments to be fed through to rgbif functions
	# Allow for user to limit query to particular taxon groups
	# Overcome harcoded 99,999 hard coded facet limit from rgbif
	# Sys.sleep might stop the function failing to return when processing data for large countries/with lots of info (e.g. Colombia in the current example)?
	# Retrieves list for one country, but can be applied over a list of country names
	getSppLists <- function (country = country, ...) {
	# Establish the minimum number of records to examine in order to capture all phyla and orders across all countries (might need to restrict this approach to the smaller countries local to a target location)
	cntFacet <- rgbif::count_facet(by='country', countries = country, removezeros = TRUE)
	# create list of all keys (and record counts) for taxa within listed countries
	# limit = 0 only refers to record downloads, not to information inspected on GBIF; facetLimit refers to the number of rows returned by 'group by' facet clause
	out1 <- tryCatch(rgbif::occ_search(country = country, # sum(cntFacet$V1) -- sum not really needed since change from running this for all countries at once, but leave in for now
	facet=c("speciesKey"), facetMincount = sum(cntFacet$V1), limit = 0, speciesKey.facetLimit = 99999), # might not be not enough for large countries
	error = function(err) NULL) # if it fails, return NULL. Means that when applied across countries the whole process will not fail
	if (is.null(out1)) {NULL} # move on (e.g. to next country in list is being used with an apply function) if this step has failed
	else {
	# need to use speciesKey return in order to get the names and higher taxon information using rgbif::name_usage
	out2 <- as.data.frame(out1$facets$speciesKey)
	out3 <- list() # empty list for all the
	out3 <- tryCatch(lapply(out2[,'name'], function(x) {tmp <- rgbif::name_usage(key = x, rank = c("phylum", "order"))
	if("order" %in% colnames(tmp$data) & "phylum" %in% colnames(tmp$data)) {tmp}
	else {}
	}), error = function(err) NULL) # get higher taxon information for later filtering
	if (is.null(out3)) {NULL} # move on (e.g. to next country in list is being used with an apply function) if this step has failed
	else { # extract and simplify structure of data returned
	out4 <- do.call(rbind, lapply(lapply(out3, '[[', 'data'), function(x) as.data.frame(x[,c('scientificName','phylum','order')]))) # works even if list out4 has NULL elements
	out4$country <- rep(attr(out1, "args")$country, nrow(out4)) # add in country information to list item
	print(paste("Retrieved and processed:", unique(out4$country))) # add country info in terms of ISO 2 code
	return(out4) # return dataframe as long as no errors are picked up from GBIF
	}
	}
	} # END

view raw retrieve_GBIF_checklists.R hosted with ❤ by GitHub