Search

Findability: Dataset Markup


Recipe metadata

identifier: RX.X

version: v0.1

Difficulty level

Reading Time

10 minutes

Recipe Type

Guidance

Executable Code

No

Intended Audience

Software Developers

Data Scientists


Main Objectives

The main purpose of this recipe is:

To embed Schema.org markup in a web page representing a dataset.


Graphical Overview of the FAIRification Recipe Objectives

graph TD A(HTML page):::box --> B{Page
just about
one dataset?}:::box B -->|YES| C(Create markup for dataset):::box B -->|NO| F(Use DataCatalog Recipe):::box C --> D(Markup Template):::box D -->|Embed template in website| E(fa:fa-search fa:fa-cog fa:fa-fighter-jet Schema.org augmented HTML page):::box classDef box font-family:avenir,font-size:14px,fill:#2a9fc9,stroke:#222,color:#fff,stroke-width:1px linkStyle 0,1,2,3,4 stroke:#2a9fc9,stroke-width:1px,color:#2a9fc9,font-family:avenir;

Capability & Maturity Table

Capability Initial Maturity Level Final Maturity Level
Findability minimal repeatable
Interoperability minimal

Method

We will outline the steps for marking up a page in your site that is about a specific dataset that you publish. The resulting markup will be compliant with both Google's Dataset markup guidelines and the Bioschemas Dataset Profile. The resulting webpage will be indexable by the major search engines and should eventually appear in Google's Dataset Search Tool.

We will use UniProtKB as an example for this recipe.

  1. Identify the page in your site about a specific dataset, e.g. https://www.uniprot.org/uniprot/

  2. Open the Bioschemas Generator

    1. Select Dataset from the Bioschemas Profile dropdown

    2. Enter the URL of the page in URL box, e.g. https://www.uniprot.org/uniprot/

    3. Click on the Show Form button

Bioschemas Generator start screen

  1. Complete the profile form with the data relevant for your page. Once completed, click on the Generate Markup button

    • You should complete all Minimum properties and as many Recommended properties as possible. You can show/hide properties using the Additional Properties buttons.
    • Where possible you should link to other resources. The Bioschemas Generator does not make this as simple as it could, but you can do it in step 5 once you have generated your markup, e.g. our dataset will link to a page with DataCatalog markup in rather than repeating all the properties for now we will just enter a url and no other properties
    • The form defaults to the data type with the first alphabetical character, e.g. for identifier this defaults to PropertyValue but Text or URL will be more appropriate in most cases
    • The right side of the screen gives examples for properties, where these have been provided by the Bioschemas profile authors. Click on the Show button to see the example for a specific property. Click on Minimum, Recommended, or Optional to expand/contract the section and see the properties contained at that marginality level

    Bioschemas Generator Dataset profile form

  2. You will now see the generated markup in JSON-LD format. You can click on the Microdata and RDFa tabs to see the same content rendered in the different formats. However, we recommend the use of JSON-LD. For our UniProtKB example, we get the following markup

    <script type="application/ld+json" >
    {
      "@context": "http://schema.org",
      "@id": "https://www.uniprot.org/uniprot/",
      "@type": "Dataset",
      "citation": [
        {
          "@id": "https://doi.org/10.1093/nar/gky1049",
          "@type": "CreativeWork"
        }
      ],
      "creator": [
        {
          "@context": "http://schema.org",
          "@type": "Organization",
          "dct:conformsTo": "https://bioschemas.org/profiles/Organization/0.2-DRAFT-2019_07_19",
          "description": "The mission of UniProt is to provide the scientific community with a comprehensive, high quality and freely accessible resource of protein sequence and functional information. ",
          "name": "UniProt Consortium"
        }
      ],
      "dct:conformsTo": "https://bioschemas.org/profiles/Dataset/0.3-RELEASE-2019_06_14",
      "description": "The UniProt Knowledgebase (UniProtKB) is the central hub for the collection of functional information on proteins, with accurate, consistent and rich annotation. In addition to capturing the core data mandatory for each UniProtKB entry (mainly, the amino acid sequence, protein name or description, taxonomic data and citation information), as much annotation information as possible is added.",
      "distribution": {
        "@id": "https://www.uniprot.org/downloads#uniprotkblink",
        "@type": "DataDownload"
      },
      "identifier": [
        "https://www.uniprot.org/uniprot/"
      ],
      "includedInDataCatalog": [
        {
          "@context": "http://schema.org",
          "@type": "DataCatalog",
          "dct:conformsTo": "https://bioschemas.org/profiles/DataCatalog/0.3-RELEASE-2019_07_01",
          "description": "",
          "keywords": [],
          "name": "",
          "url": "https://uniprot.org"
        }
      ],
      "keywords": [
        "Protein",
        "Protein annotation"
      ],
      "license": "http://creativecommons.org/licenses/by/4.0/",
      "name": "UniProtKB",
      "url": "https://www.uniprot.org/uniprot/"
    }
    </script >
    
  3. Download or copy and paste the generated markup

  4. Make adjustments for any bits that could not be properly entered through the form.

    For example, for our generated markup we would change the includedInDataCatalog so that it provides a direct link rather than repeating the properties. We would replace

    "includedInDataCatalog": [
        {
          "@context": "http://schema.org",
          "@type": "DataCatalog",
          "dct:conformsTo": "https://bioschemas.org/profiles/DataCatalog/0.3-RELEASE-2019_07_01",
          "description": "",
          "keywords": [],
          "name": "",
          "url": "https://uniprot.org"
        }
      ],
    

    with

    "includedInDataCatalog": {
       "@type": "DataCatalog",
       "@id": "https://uniprot.org"
     },
    

    You can test that your JSON-LD is valid syntax, and visualise your markup using the JSON-LD Playground.

  5. Once you are happy with your markup, include the JSON-LD, script tags and all, at the bottom of your HTML page template. Make sure that this is before the closing </html> tag

  6. If you have multiple datasets released through your site, then you should make a template for your datasets. In your template you should replace the values in your markup that will change from dataset to dataset with variables. Your web page templating system will replace the variables with values from your database. For example, the follow snippet uses variables of the form %%%PAGEURL%%%

    <script type="application/ld+json">
    {
      "@context": "http://schema.org",
      "@id": "%%%PAGEURL%%%",
      "@type": "Dataset",
      "citation": [
        {
          "@id": "%%%DOI%%%",
          "@type": "CreativeWork"
        }
      ],
      "creator": [
        {
          "@context": "http://schema.org",
          "@type": "Organization",
          "dct:conformsTo": "https://bioschemas.org/profiles/Organization/0.2-DRAFT-2019_07_19",
          "description": "The mission of UniProt is to provide the scientific community with a comprehensive, high quality and freely accessible resource of protein sequence and functional information. ",
          "name": "UniProt Consortium"
        }
        ...
      ]
    }
    </script>
    

Your site should now generate dataset pages with embedded markup.

Once you have deployed this on your web server, you can test it with the Bioschemas Validator which scrapes the markup from your page and allows you to test it against various Bioschemas profiles[1](#bioschemas-validator).


FAIRification Objectives, Inputs and Outputs

Actions.Objectives.Tasks Input Output
text annotation Bioschemas annotated text
validation schema.org report

Table of Data Standards

Data Formats Terminologies Models
JSON-LD Bioschemas RDF
HTML

Authors:

Name Affiliation orcid CrediT role
Alasdair Gray Bioschemas Community Lead / Heriot-Watt Unviersity / ELIXIR-UK 0000-0002-5711-4872 Writing - Original Draft
Leyla Garcia Bioschemas Community / ZB MED Information Centre for life sciences, Knowledge Management Group 0000-0003-3986-0510 External review

Footnotes:

1: The Bioschemas Validator is currently in an early alpha release and does not include all the profiles.


License:

This page is released under the Creative Commons 4.0 BY license.