Create Lightweight Schema.org Descriptions of Data.

dataspice

The goal of dataspice is to make it easier for researchers to create basic, lightweight, and concise metadata files for their datasets by editing the kind of files they’re probably most familiar with: CSVs. To spice up their data with a dash of metadata. These metadata files can then be used to:

• Make useful information available during analysis.
• Create a helpful dataset README webpage for your data similar to how pkgdown creates websites for R packages.
• Produce more complex metadata formats for richer description of your datasets and to aid dataset discovery.

Metadata fields are based on Schema.org/Dataset and other metadata standards and represent a lowest common denominator which means converting between formats should be relatively straightforward.

Example

An basic example repository for demonstrating what using dataspice might look like can be found at https://github.com/amoeba/dataspice-example. From there, you can also check out a preview of the HTML dataspice generates at https://amoeba.github.io/dataspice-example and how Google sees it at https://search.google.com/test/rich-results?url=https%3A%2F%2Famoeba.github.io%2Fdataspice-example%2F.

A much more detailed example has been created by Anna Krystalli at https://annakrystalli.me/dataspice-tutorial/ (GitHub repo).

Installation

You can install the latest version from CRAN:

install.packages("dataspice")

Workflow

create_spice()
# Then fill in template CSV files, more on this below
write_spice()
build_site() # Optional

Create spice

create_spice() creates template metadata spreadsheets in a folder (by default created in the data folder in the current working directory).

The template files are:

• biblio.csv - for title, abstract, spatial and temporal coverage, etc.
• creators.csv - for data authors
• attributes.csv - explains each of the variables in the dataset
• access.csv - for files, file types, and download URLs (if appropriate)

Fill in templates

The user needs to fill in the details of the four template files. These csv files can be directly modified, or they can be edited using either the associated helper function and/or Shiny app.

Helper functions

• prep_attributes() populates the fileName and variableName columns of the attributes.csv file using the header row of the data files.

• prep_access() populates the fileName, name and encodingFormat columns of the access.csv file from the files in the folder containing the data.

To see an example of how prep_attributes() works, load the data files that ship with the package:

data_files <- list.files(system.file("example-dataset/", package = "dataspice"),
  pattern = ".csv",
  full.names = TRUE
)

This function assumes that the metadata templates are in a folder called metadata within a data folder.

attributes_path <- file.path("data", "metadata", "attributes.csv")

Using purrr::map(), this function can be applied over multiple files to populate the header names

data_files %>%
  purrr::map(~ prep_attributes(.x, attributes_path),
    attributes_path = attributes_path
  )

The output of prep_attributes() has the first two columns filled out:

Shiny helper apps

Each of the metadata templates can be edited interactively using a Shiny app:

• edit_attributes() opens a Shiny app that can be used to edit attributes.csv. The Shiny app displays the current attributes table and lets the user fill in an informative description and units (e.g. meters, hectares, etc.) for each variable.
• edit_access() opens an editable version of access.csv
• edit_creators() opens an editable version of creators.csv
• edit_biblio() opens an editable version of biblio.csv

Remember to click on Save when finished editing.

Completed metadata files

The first few rows of the completed metadata tables in this example will look like this:

access.csv has one row for each file

attributes.csv has one row for each variable in each file

biblio.csv is one row containing descriptors including spatial and temporal coverage

creators.csv has one row for each of the dataset authors

Save JSON-LD file

write_spice() generates a json-ld file (“linked data”) to aid in dataset discovery, creation of more extensive metadata (e.g. EML), and creating a website.

Here’s a view of the dataspice.json file of the example data:

Build website

• build_site() creates a bare-bones index.html file in the repository docs folder with a simple view of the dataset with the metadata and an interactive map. For example, this repository results in this website

Convert to EML

The metadata fields dataspice uses are based largely on their compatibility with terms from Schema.org. However, dataspice metadata can be converted to Ecological Metadata Language (EML), a much richer schema. The conversion isn’t perfect but dataspice will do its best to convert your dataspice metadata to EML:

library(dataspice)

# Load an example dataspice JSON that comes installed with the package
spice <- system.file(
  "examples", "annual-escapement.json",
  package = "dataspice"
)

# Convert it to EML
eml_doc <- spice_to_eml(spice)
#> Warning: variableMeasured not crosswalked to EML because we don't have enough
#> information. Use `crosswalk_variables` to create the start of an EML attributes
#> table. See ?crosswalk_variables for help.
#> You might want to run EML::eml_validate on the result at this point and fix what validations errors are produced. You will commonly need to set `packageId`, `system`, and provide `attributeList` elements for each `dataTable`.

You may receive warnings depending on which dataspice fields you filled in and this process will very likely produce an invalid EML record which is totally fine:

library(EML)
#> 
#> Attaching package: 'EML'
#> The following object is masked from 'package:magrittr':
#> 
#>     set_attributes

eml_validate(eml_doc)
#> [1] FALSE
#> attr(,"errors")
#> [1] "Element '{https://eml.ecoinformatics.org/eml-2.2.0}eml': The attribute 'packageId' is required but missing."                                  
#> [2] "Element '{https://eml.ecoinformatics.org/eml-2.2.0}eml': The attribute 'system' is required but missing."                                     
#> [3] "Element 'dataTable': Missing child element(s). Expected is one of ( physical, coverage, methods, additionalInfo, annotation, attributeList )."
#> [4] "Element 'dataTable': Missing child element(s). Expected is one of ( physical, coverage, methods, additionalInfo, annotation, attributeList )."
#> [5] "Element 'dataTable': Missing child element(s). Expected is one of ( physical, coverage, methods, additionalInfo, annotation, attributeList )."

This is because some fields in dataspice store information in different structures and because EML requires many fields that dataspice doesn’t have fields for. At this point, you should look over the validation errors produced by EML::eml_validate and fix those. Note that this will likely require familiarity with the EML Schema and the EML package.

Once you’re done, you can write out an EML XML file:

out_path <- tempfile()
write_eml(eml_doc, out_path)
#> NULL

Convert from EML

Like converting dataspice to EML, we can convert an existing EML record to a set of dataspice metadata tables which we can then work from within dataspice:

library(EML)

eml_path <- system.file("example-dataset/broodTable_metadata.xml", package = "dataspice")
eml <- read_eml(eml_path)

# Creates four CSVs files in the `data/metadata` directory
my_spice <- eml_to_spice(eml, "data/metadata")

Resources

A few existing tools & data standards to help users in specific domains:

• Darwin Core
• Ecological Metadata Language (EML) (& EML)
• ISO 19115 - Geographic Information Metadata
• ISO 19139 - Geographic Info Metadata XML schema
• Minimum Information for Biological and Biomedical Investigations (MIBBI)

…And others indexed in Fairsharing.org & the RDA metadata directory.

Code of Conduct

Please note that this package is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

Contributors

This package was developed at rOpenSci’s 2018 unconf by (in alphabetical order):

• Carl Boettiger
• Scott Chamberlain
• Auriel Fournier
• Kelly Hondula
• Anna Krystalli
• Bryce Mecum
• Maëlle Salmon
• Kate Webbink
• Kara Woo
• Irene Steves.