An overview of ArctosR

ArctosR is a package designed to download data from Arctos, format these data to make it easier for users to read and relate content, and save results in various formats. This vignette provides an overview of the basic usage of this package.

Basics

Types of data in Arctos

Arctos contains specimen records and diverse types of data associated with them (e.g., measurements, locality coordinates and descriptions, tissue samples available, etc.). For a full description of Arctos and its data visit its website.

Queries and responses in ArctosR

ArctosR abstracts interacting with the Arctos API via objects. These objects can be manipulated with a set of using friendly functions for most tasks, or created and manipulated directly using builder functions.

Query objects

A query in ArctosR is a collection of searches by a user of the package for a specific task. This can be: (1) a simple search, such as a one-off search for specimens of a species held by some museum; (2) a search that requires the concatenation of multiple responses, such as one that requests more records than Arctos can provide in a single response; (3) or a complex search that uses requests for one set of search terms, then uses those responses as part of another request.

Response objects

A response then is an object that stores the contents of one response from Arctos back to ArctosR. Each response has associated metadata, such as search terms and time-stamp, and content, such as a table of records matching those search terms.

Metadata

Detailed metadata about each response is also saved in the user’s query for the purposes of documentation and reproducibility. This metadata is saved along with downloaded records as a JSON file.

Making queries to Arctos

Make sure to load the package:

library(ArctosR)

A basic query

ArctosR provides a single function, get_records, to search for records in the Arctos database. In order to start building a search, we have to find out the possible query parameters we can use to search. These can be found with the get_query_parameters function, which returns a dataframe of all query parameters used by Arctos. The names in the obj_name column are what are used as parameters to the get_records function.

# run the function and store results in an object
query_params <- get_query_parameters()

# checking the dataframe obtained (showing only 6 rows and 3 columns)
query_params[1:6,1:3]
#>                    display              obj_name category
#> 1            Verbatim Date         verbatim_date    event
#> 2     Chronological Extent  chronological_extent    event
#> 3 Collecting Event Remarks    coll_event_remarks    event
#> 4        Collecting Source     collecting_source    event
#> 5        Collecting Method     collecting_method    event
#> 6    Collecting Event Name collecting_event_name    event
#>             display                obj_name category
#> 1     Verbatim Date           verbatim_date    event
#> 2 Collecting Method       collecting_method    event
#> 3 Collecting Source       collecting_source    event
#> 4        Ended Date              ended_date    event
#> 5  Event Attributes evtAttributeSearchTable    event
#> 6           Habitat                 habitat    event

For this basic query, we can use guid_prefix, whose description can be found by listing the row of the query parameter dataframe:

# checking row 37 in the dataframe
query_params[37,1:5]
#>     display            obj_name       category subcategory
#> 37 Taxonomy taxonomySearchTable identification    taxonomy
#>                                                                                                                                                                 description
#> 37 Taxonomic name in any taxonomy source. Taxonomy used in identifications: Search VALUE OF (and specify an operator), optionally AS RANK, optionally ACCORDING TO SOURCE.
#>       display    obj_name   category subcategory
#> 37 Collection guid_prefix identifier       basic
#>                                                                description
#> 37 Collection responsible for the record. Turning this off will break most
#>    forms.

We will also use genus and species whose description can be found at rows 28 and 21, respectively:

query_params[28,1:5]
#>    display obj_name       category subcategory
#> 28 Kingdom  kingdom identification  curatorial
#>                                                 description
#> 28 Kingdom as provided in collection's preferred Source(s).
#>    display obj_name       category subcategory
#> 28   Genus    genus identification  curatorial 
#>                                               description
#> 28 Genus as provided in collection's preferred Source(s).

query_params[23,1:5]
#>      display  obj_name       category subcategory
#> 23 Subfamily subfamily identification  curatorial
#>                                                   description
#> 23 Subfamily as provided in collection's preferred Source(s).
#>    display obj_name       category subcategory 
#> 23 Species  species identification  curatorial
#>                                                            description
#> 23 Species (binomial) as provided in collection's preferred Source(s).

Now that we have a set of parameters to use, we can pass them to the get_records function and send our request to Arctos. This will return a query, which bundles our search parameters with the returned data from Arctos.

query <- get_records(guid_prefix = "MSB:Mamm", genus = "Canis", 
                     species = "lupus")

This simple search returned the default (core) columns provided by Arctos. There is a a lot more information that can be requested from Arctos, see below for an example of how to do it.

Requesting other columns than the core set

By default, the columns returned by Arctos are all of those with the category core. These columns can be listed as follows:

result_params <- get_result_parameters()
result_params[result_params$category == 'core',1:2]
#>                      display                      obj_name
#> 1  GUID (DarwinCore Triplet)                          guid
#> 7              Identified As               scientific_name
#> 43          Asserted Country                       country
#> 44   Asserted State/Province                    state_prov
#> 53         Specific Locality                 spec_locality
#> 59             Verbatim Date                 verbatim_date
#> 71          Decimal Latitude                       dec_lat
#> 72         Decimal Longitude                      dec_long
#> 73      Coordinate Error (m) coordinateuncertaintyinmeters
#>                      display                      obj_name
#> 1  GUID (DarwinCore Triplet)                          guid
#> 7              Identified As               scientific_name
#> 43          Asserted Country                       country
#> 44   Asserted State/Province                    state_prov
#> 51         Specific Locality                 spec_locality
#> 57             Verbatim Date                 verbatim_date
#> 69          Decimal Latitude                       dec_lat
#> 70         Decimal Longitude                      dec_long
#> 71      Coordinate Error (m) coordinateuncertaintyinmeters

Additional columns can be requested by passing a vector of result parameters to the get_records function in the columns parameter like so:

# making a list of additional columns to get (see get_query_parameters())
add_cols <- list("guid", "scientific_name", "relatedcatalogeditems", "collectors",
                 "state_prov", "spec_locality", "dec_lat", "dec_long", 
                 "verbatim_date", "examined_for", "detected", "not_detected")

# getting records with additional columns
query <- get_records(guid_prefix = "MSB:Mamm", genus = "Canis", 
                     species = "lupus", columns = add_cols)

Requesting columns that are tables

Certain result parameters (columns) in Arctos are entire tables associated to a single specimen record. For instance, partdetail, which links to the attributes of each part listed in parts associated with a specimen. These can be requested just like any other result parameter. The information in these complex columns is obtained in JSON format, but can be expanded into dataframes of their own with the function expand_column.

# defining the columns to be obtained
some_cols <- list("guid", "parts", "partdetail")

# performing the query
query <- get_records(guid_prefix = "MSB:Mamm", genus = "Canis", 
                     species = "lupus", columns = some_cols)

See an example of expanding the columns in the section Expanding columns

Requesting all records

By default, get_records avoids requesting all records for a query unless otherwise asked. By passing the parameter all_records = TRUE to get_records, the user can request that ArctosR make multiple requests until all records for a given query are downloaded.

get_record_count(guid_prefix = "MSB:Mamm", genus = "Canis", species = "lupus")
#> [1] 1694

query <- get_records(guid_prefix = "MSB:Mamm", genus = "Canis", 
                     species = "lupus", all_records = TRUE)

Downloading and using data from Arctos

We have gone through the basic functionality of ArctosR. Below you can find an example of using the package to get, process, explore, and save data from Arctos.

Get and view data

# a list of columns to download with the query
my_cols <- list("guid", "scientific_name", "parts", "collectors", "state_prov", 
                "spec_locality", "dec_lat", "dec_long", "verbatim_date",
                "partdetail")

# download records
query <- get_records(guid_prefix = "MSB:Mamm", genus = "Canis", 
                     species = "lupus", columns = my_cols)

# getting only the dataframe of data
msb_wolves <- response_data(query)

Expanding columns

The column partdetail contains nested dataframes associated to each record. To explore these data fully, we can process the information and view it as a table.

# process the information in partdetail into sub-dataframes
expand_column(query, "partdetail")

Saving data

ArctosR offers multiple options to save the data obtained from Arctos. See below for examples of how to do it.

Saving as a flat CSV file

save_response_csv(query, "msb_wolves.csv")

Saving an expanded CSV

save_response_csv(query, "msb_wolves.csv", expanded = TRUE)

Saving as an RDS

save_response_rds(query, "msb_wolves.rds")

Saving query metadata

save_response_csv(query, "msb_wolves2.csv", with_metadata = TRUE)