Thank you to everyone who responded to our request for feedback on our metadata draft! We really appreciated your comments and thoughts. Below is a summary of how we are incorporating comments and changes based on feedback from an internal survey, external survey, and our governance committee.
You can view the revised draft and our comments here.
Changes we made
We made a handful of changes based on the feedback:
- Removed field 27-Spatial/geo coverage as we concluded this was more confusing than not and that publishing and quality checks would serve to ensure completeness.
- Changed the fields 12 and 13 to “Required - Private” per feedback from our governance committee who felt that they created unnecessary confusion given that they supported internal management versus understanding the dataset. For example, public users would only see datasets marked as public access level: public. As a result, providing this as a public field doesn’t add value and may create confusion.
- Incorporated some changes to the drop down menus and what we would include.
Guidance we’ll incorporate
A few of you had suggestions for new field elements or information to include. Since our research noted that “compliance” is an issue, we are very focused on making the fields easy to populate while recognizing that good metadata is essential.
To balance ease of use with comprehensiveness, we’ll create templates and guidance for completing all of the field elements, but with a focus on:
- 1-Title
- 2-Description
- 5-Data dictionary
- 25-Data notes
- 26-Related documents
With the goal of addressing questions about:
- Data collection methods and business processes generating the data
- Units of analysis or suggested analysis
- Primary keys and other important data fields
And of course, we will monitor this in the next year and see if we need to add additional field elements or improve our guidance.
Ease of completion
We heard from a few folks concerned about how much effort it will take to complete the metadata. In addition, to the guidance we’ll develop, we worked hard to minimize manual input/writing.
Of the 15 required elements, 10 are either auto-generated or drop downs, which simplifies completion. Another 2 are simple text - that is they should be simple and straightforward to complete. We’ll provide detailed guidance on creating the title and description which are open text and the data dictionary, which will require a mix of strategies to complete. The ratio is similar for the conditionally required elements - 5 of the 7 are auto-generated or drop downs.
Table: Metadata field by type of field
| Field Type | Required | Conditionally required | Optional | Total |
|---|---|---|---|---|
| Auto-generated | 4 | 3 | 7 | |
| Drop down | 6 | 2 | 8 | |
| Simple text | 2 | 1 | 2 | 5 |
| Open text | 2 | 1 | 1 | 4 |
| Other | 1 | 1 | 2 | |
| Total | 15 | 7 | 4 | 26 |
We hope the combination of clear guidance and use of auto-generation and drop downs will help keep things manageable.
Thank you again to everyone who contributed feedback! We really appreciate it and continue to look forward to working with everyone on improving the delivery and quality of data through our open data portal.
Your Open Data Team,
Joy & Jason