The ThinkData Works Catalog provides a very flexible way to add metadata to a dataset, letting you add as many custom fields as you want to any dataset in the catalog.
Sometimes, however, it's important to create a standard set of metadata that you'd like all datasets from a specific source to adhere to. This is commonly called a "taxonomy", or "ontology". Put simply, it's a collection of similar metadata that you'd like to attach to many datasets.
As an example, you may have a team who requires the same metadata on all the datasets they use. In order to use the dataset effectively, they need to know:
- License cost
- License pdf
- Point of contact
- Acceptable uses
Rather than add all these as separate fields onto every dataset in the source, a user may build a "metadata template" that they may apply to many datasets, ensuring the same fields are present and, by extension, standardizing how information is managed in the catalog.
To build a metadata template, go to the "Metadata" tab in the sidebar. From there, navigate to the "metadata templates" section. Select "Create template" to begin.
Building a metadata template is similar to adding custom fields onto a dataset. You begin by adding whatever fields you want to the template. For the above example, I'd add the "Data cost" field, an attachment field, the "Point of contact" field, and a list field.
For templated metadata you may provide a default value, or leave the values blank so they can be filled in on a dataset by dataset basis. For a list type field, you may include many options, one of which may be used by the person adding the metadata. For our "acceptable uses" field, we'd potentially include the options "all uses, non-commercial use, limited use". You may also set all fields as required or optional. Finish by giving your template a name. This can be anything you like, but try to make sure it's easy to understand. "Data Team Metadata Taxonomy" is a stronger name than "Template 1"!
Once you have created the template you want, select "Save". You will see it in the list of your other metadata templates.
At this point you may add the template to datasets. To do so, open a dataset and select edit, then scroll down to the metadata section and select the "Add metadata" button. From within the popover you will be able to select the template you just created.