Skip to content

Search for Research Data

Search the ANDS Site

Search
Search
http://www.ands.org.au http://www.ands.org.au

Metadata for impact: Make RIF-CS work for you

Share
Share

The ANDS RIF-CS metadata requirements show that some elements are required (or mandatory) while others are recommended (or optional). Creating metadata descriptions involves some effort, so how should you decide which recommended elements to include in your data descriptions?

A good way to think about this is to consider what your institution wants to achieve by publishing data via Research Data Australia (RDA) and how you expect people will search for and reuse your data. 
Some common institutional goals, with examples of associated RIF-CS encoding, are provided below.

For more information about maximising the impact of your data:

1. We want to highlight our open data. The transparency of our research is important to our reputation. Which RIF-CS elements are important to us?

As open data is citable and reusable it brings many benefits to individuals, institutions and the broader research community.  The 'openness' of data is defined by several characteristics, so describing this requires more than one element. Providing content for these elements will not only provide users with clarity around access and reuse conditions, it will also ensure the record is prominently displayed in RDA.

Provide content for these RIF-CS elements:

  • collection:rights:type="accessRights" - choose 'open' from the suggested vocabulary to indicate that the data is publicly accessible online.  RDA users are able to limit their search to open data.
  • collection:rights:type="licence" - an open licence such as CC-BY will ensure you are attributed when the data is reused. If no licence is assigned, reuse of the data is not permitted, so be sure to assign an appropriate licence and describe it in this element.
  • collection:location:address:type="electronic" - include the @target attribute that indicates the data set is available for download with a link provided to do so. The link may trigger a direct download of the data or link to a metadata landing page from which the data may be downloaded. You can also provide additional information such as title, media type, byte size and notes.

A defining characteristic of open data is that it is well described, so include as much rich information as possible in collection records to describe what the data is, how it was collected and for what purpose.

Example 1.1: xml encoding of the rights element incorporating licence and accessRights.
<rights>
   <rightsStatement rightsUri="http://sydney.edu.au/copyright/">Copyright, The University of Sydney, 2014</rightsStatement>
   <licence type="CC-BY" rightsUri="http://creativecommons.org/"</licence>
   <accessRights type="Open" rightsUri="http://sydney.edu.au/access/">In accordance with University policy, this resource may be downloaded free of charge</accessRights>
</rights>
Example 1.2: Registry view of the encoding

Registry accessRights

Example 1.3 xml encoding of the location element showing electronic address for a sample (fictitious) record

<location>
 <address>
  <electronic type="url" target="directDownload">
   <value>http://www.ga.gov.au/corporate_data/64739/64739_sh50-04_kml.zip</value>
  <title>Youanmi Map Series</title>
  <mediaType>application/zip/kml</mediaType>
  <byteSize>5 MB</byteSize>
  </electronic>
 </address>
< /location>

Example 1.4 Registry view of the encoding
Registry electronic address
 2. We want citation metrics for our data - the same as our publications. It is another way to demonstrate the impact of our research and how me might identify new collaborators. Which RIF-CS elements are needed to enable this?

ANDS has established a process for harvesting records from a data source in RDA to the Thomson Reuters Data Citation Index. This will enable metrics for citation of data to be counted in the same way as publications. As similar indices emerge, ANDS will seek to ensure records in RDA are included. 

Provide content for these RIF-CS elements:

  • collection:citationInfo:citationMetadata - enables you to describe, in a structured way, how the data should be cited. Data that is cited in a consistent (and machine readable) way is readily identifiable by citation indexing services such as Thomson Reuters Data Citation Index.
  • collection:identifier - if possible, assign a DOI (Digital Object Identifier) to your data. DOIs uniquely identify a dataset and are considered 'best practice' for the accurate capture of citation metrics. ANDS offers the DOI Service (Cite My Data).  It is available to publically funded research organisations as a machine to machine or manual service.

If you'd like the records in your data source to be harvested to the Data Citation Index take a look at the DCI encoding notes  to ensure your records are ready to be harvested and can start accruing metrics.
See also
3. Linking data to related publications.

Example 2.1: xml encoding of citationMetadata, with DOI as the identifier, for a sample (fictitious) record

<citationInfo>
    <citationMetadata>
    <identifier type="doi">10.4225/13/50BBFCFE08A12</identifier>
    <title>Surface water run-off measurements in the City of Salisbury, South Australia during the period June 2012 to December 2012</title>
    <version>1</version>
    <placePublished/>
    <publisher>The University of South Australia</publisher>
    <url/>
    <context/>
    <contributor seq="1">
    <namePart type="family">Oliver</namePart>
    <namePart type="given">R</namePart>
    </contributor>
    <contributor seq="2">
    <namePart type="family">Myers</namePart>
    <namePart type="given">B</namePart>
    </contributor>
    <date type="publicationDate">2013</date>
</citationMetadata>

 Example 2.2: Registry view of the encoding

Registry citation

3. We want to be able to link our published data to related publications.  It may help drive up the citation count for our publications.  How can we show these linkages in RIF-CS?

While it is not possible to create a registry object (or separate RIF-CS record) to describe a publication in Research Data Australia, it is possible to provide information about a publication using the relatedInfo element in the relevant collection record.  Providing this information can bring multiple benefits including:

  1. linking data to related publications can provide rich contextual information to support data reuse;
  2. this information can be used to link data records in the Data Citation Index to related publications in the Web of Science citation indices, providing enhanced discovery of both data and publications. More information about Data Citation Index.
  3. evidence suggests that publications providing access to related data are more likely to be cited.

Provide content for this RIF-CS element:

  • collection:relatedInfo:type="publication" - enables you to describe a publication associated with a data collection.  Where possible, include the DOI for the publication and provide the full citation as a note.
Example 3.1: xml encoding of relatedInfo to link a data collection to a related publication (fictitious example).

<relatedInfo type="publication">
   <title>Surface water quality in the City of Salisbury, South Australia</title>
   <identifier type="doi">10.4225/08/53EC60AB0DD1B</identifier<relation type="isCitedBy"/>
   <notes> Myers, B., Oliver, R. & Pezzaniti, D. (2013) Surface water quality in the City of Salisbury, South Australia.  Australian Journal of Water Quality, vol.12, no.5, pp.4-9 </notes>
</relatedInfo>

Example 3.2: Registry view of the encoding

Registry related publicaiton

4. We want people to know they can access and use our data via a service. Which RIF-CS elements should we be looking at? 

In some cases, data described in RDA is accessible via a service. The service may enable a user to download the data being described, or it may allow a user to "do something" with the data. For example, create a visualisation of the data. Two options exist to describe a service related to a data collection:

  1. Create a separate registry object (or RIF-CS record) to describe the service and link that record to the relevant collection record(s) using the relatedObject element.  This option will give greater exposure to the data service by ensuring it is readily discoverable in RDA.  This option is preferred where the description is being provided by the service owner or maintainer.
  2. Provide information about the service within a collection record using the relatedInfo element.  This option provides sufficient information for a user to access a collection via a service without providing a description of the service itself. Use this option where option 1 is not feasible or the service is owned and maintained by another organisation.

For Option 1, follow the best practice guidance for creating service records.

For Option 2, provide content for this RIF-CS element:

  • collection:relatedInfo:type="service" - enables you to describe a service related to a collection.  Include a URI for the service and where possible, a direct link to the data described.
Example 4.1: xml encoding for Option 2 (relatedInfo) for a sample (fictitious) record

<relatedInfo type="service">
 <title>Marine Virtual Laboratory Information System</title>
 <identifier type="uri">http://marvlis.aodn.org.au/marvlis</identifier>
 <relation type="supports">
 <url>http://marvlis.aodn.org.au/marvlis/ACQ_SurfPlt/MAPWaterTemps.png</url>
 </relation>
 <notes>Data visualisation</notes>
</relatedInfo>

Example 4.2: Registry view of the encoding

Registry related service