Getting Started - How You Can Publish Your Metadata

There is no one-size-fits-all approach to metadata. This guide is an overview of the steps in the metadata creation process from initial planning through publication in a metadata registry or repository. Even if your ultimate goal is not to publish your metadata, the initial steps below will still be relevant for planning your in-house metadata. 

You will need to develop a metadata template, that is, a structure to organize your metadata that is based on standards and has been converted into an electronic format. You will then need to populate that template with your metadata, and then publish it to the outlet of your choice. This can be divided into the following steps:

1. Establish collaborations

To prepare for publishing your metadata, collaborate with the science and technical professionals in your organization. Your metadata project will benefit from both types of expertise, so initiate the collaboration early.

2. Consider metadata in early project planning

A common problem in writing metadata is not having the information required to actually fill in the metadata. To prevent this pitfall, plan the metadata before a new project begins. Developing your plan before any type of data collection or instrument deployment has occurred will assure both thoroughness and relevance.

Your plan should include how the metadata will be created, updated, and disseminated. Recording metadata accurately as the project proceeds is much easier than organizing it after it has been collected. For example, the metadata creation component should include a protocol for field metadata collection including media, methods, and information to be collected. Knowing that information to be collected might include instrument settings or environmental conditions will make sure those metadata are not overlooked.

3. Evaluate your data and how it will be used

Get to know your data (or get to know someone who knows your data). Keep track of the questions that arise as you become acquainted with the data. Consider how you will use the data and what you will need to know to use it.

By creating metadata, you help promote the long-term use of the data set you are describing. When planning a metadata approach, think about how future users might view your metadata.

For example, consider a hypothetical situation ten years from now when a researcher is conducting a study that requires the blending of multiple oceanographic data sets. To do the study, the researcher must verify that your data was collected using appropriate instrumentation and that appropriate techniques were used for the post-processing (for example, calibrations, screening, algorithms, and assumptions). Your metadata should be capable of supporting such a user.  Understand your likely user group and prepare a long-term metadata strategy.

4. Select one (or more) metadata standards

Metadata standards are formal descriptions of the content and, in some cases, the format that metadata should have (See Metadata Standards for more information on finding and selecting metadata standards). Adopting an existing metadata standard can make your job easier by providing an established template and tools, which will make it easier to manage, share, and archive both your data and metadata. It is important to be aware of the standards that apply to metadata in general, and to your project in particular, and which standards are used in your domain. If a standard seems confusing, redundant, or contradictory, consult with experts to clarify inconsistencies before implementing it. See the guide on Selecting a Standard for more information.

5. Create a metadata template that fits your data

Your goal is to develop a list of metadata elements that, when completed and associated with a particular data resource, will completely describe your data and put it in context with similar projects. Examples of common metadata elements include the names of the parameters measured, the location of the measurements (generally latitude, longitude and depth), and the contact information of the data provider. Use your scientific expertise to determine how best to describe your data using the selected standards as a general framework. Some of the elements included in a standard might be mandatory. Some of the optional elements might be appropriate, while others might not. Identify if there are any conflicts with the selected standards where the description of your data doesn’t fit into the standard. You may wish to supplement the core elements from a standard with elements you think are critical to your specific data. 

Ideally, your template will facilitate the creation of metadata that makes sense for your data, adequately allows for discovery and reuse of the data, and appropriately satisfies reporting requirements. It is important to note that your set of metadata elements may evolve over time. While planning ahead is important, you might later realize that you actually need another element or that you have information that you don’t need. It is fine to adjust as needed.  

Tips for selecting metadata elements for your template

  • Review the questions you asked when you were getting to know your data. Answers to those questions will most likely help the scientific community find and use your data, so this information should be included in your metadata.
  • Look at what other projects have done. Search for their data and examine the metadata, making comparisons with your project.
  • Review the documentation associated with the relevant standards. Some standards will include sample templates.

Helpful tools

As you are exploring the standards, you will want to develop a working list of elements. Two useful tools for this are Microsoft Excel and Freemind.

Excel provides a tabular approach to your metadata template. The simple example template below is divided into two major metadata sections: Collection and Data. Your metadata will include more major sections. In this example, the first column provides the general category of metadata, the second provides the list of elements that will be included, and the third gives controlled vocabularies where appropriate.  

Freemind provides a graphical approach to your metadata template. The figure below shows the same metadata information in the table above, expressed as a Freemind map. 

check

There are a variety of other tools developed for metadata template development. Some of the more widely used metadata standards have existing templates or entry interfaces to facilitate their use. 

6. Determine your metadata format

Once the template includes metadata elements and the beginnings of controlled vocabularies, then it needs to be deployed into the project's technological infrastructure. In other words, how you capture the metadata digitally and link it to your data objects in a way that can be searched and used by humans and computers.

Storing your metadata

Choosing a format for storage of your metadata will affect long-term accessibility. One of the most versatile formats for your metadata is a comma separated value (CSV) flat text file (ASCII). By storing your metadata in a CSV ASCII file, your metadata may be deployed and translated into a variety of formats. Using the metadata example above (see the Excel and Freemind diagrams), the corresponding CSV metadata file might look like this:

"Collection_CollectionTitle","RV Melville Cruise MV0909"
"Collection_Curator","SIO Geological Data Center"
"Collection_ContactInformation","http://someurl.edu"
"Data_Filename","NavFile.txt"
"Data_Filetype","Data"
"Data_Description","This is sample metadata for text-formatted navigation data"
"Data_DataType","Navigation"

In this example system, you would generate a single text-metadata file for each data document in the system.

Because delimited text is harder for the human eye to scan and work with, entering data in an Excel spreadsheet sheet and exporting it as delimited text is useful. Extensible Markup Language (XML) is another text-based format that many standards are adopting. 

Ideally, you want to keep your metadata closely coupled, or linked, to your data object so that the metadata aren’t lost when the data are moved or distributed. Headers within files, zipped or tarred bundles that include data and metadata files, relational and other databases, and Excel workbooks with sheets for data and sheets for metadata are all approaches for doing this. If you are sharing your data with others, or archiving it for long-term preservation, it is better to avoid proprietary formats like Excel and relational database management systems in favor of text formats. Your decision will be based on the nature of your data (some data are more suited to flat files, others are more relational) and the infrastructure capabilities of your project or institution. 

7.  Implementation - capturing your metadata

Once you know how you will record your metadata, you will need to develop a plan to capture it.  For each piece of metadata, consider how it will be recorded, by whom, and when. In some cases, you can draw metadata (and data) directly from instruments. In other cases, a researcher will have to note down information. Determine the most efficient process, test it on a small scale, make adjustments as needed, and then implement it to gather your metadata.

8. Check your metadata

Once your metadata has been entered into your system, it should be reviewed for consistency, completeness, ambiguity of terms, appropriateness of title, and readability of input. See the guide on Writing Good Metadata for more information.

9. Publish

The previous steps were preparatory to the final goal of sharing your data with your research community through publication. A typical location to publish your project metadata would be a metadata clearinghouse, also called a registry or repository. The National Ocean Data Center and the Global Change Master Directory are two examples of clearinghouses. Clearinghouses permit automatic searching of metadata, so that researchers can find out that your data exist. Ideally, the metadata should be available to the international community. This will provide maximum exposure of the quality data sets that you and your organization produce. If you choose not to use a clearinghouse, you can make your metadata available to your target audience through some other means, such as a web-based portal, though we strongly recommend registration in addition to individual website creation.

After you have completed the uploading process to your location of choice, check to make sure the publication process was a success. Also, check the search capabilities of the clearinghouse to confirm that your entry will be visible or discoverable in ways that you expect. If it isn't and you don't understand why, contact the clearinghouse for help and clarification.

10. Participate in the MMI community as an experienced metadata publisher

Share your success with the marine metadata community. Although this is not part of the publication process, participating in the larger community and communicating your experiences will help improve metadata processes. You can contribute a case study about how developed your metadata and what you learned in that process, or suggest additions or changes to these guides.  

Suggested Citation

Neiswender, C., Isenor, A., Stocks, K. 2010. "Getting Started - How You Can Publish Your Metadata." In The MMI Guides: Navigating the World of Marine Metadata. http://marinemetadata.org/guides/mdataintro/gettingstarted. Accessed December 12, 2019.