Federal Data Excellence Assessment

This is an interactive, simplified version of the full rubric used to assess data products for the Federal Data Excellence Awards. This rubric can be used to assess any federal data product. Each section contains a series of “yes or no” questions that should be answered while viewing the data product being assessed. On the final screen, you will be presented with a letter grade and recommendations for how to improve the product. For the full assessment rubric that includes detailed scoring instructions and descriptive screen shots, see USAFacts.

Please enable JavaScript in your browser to complete this form.

– Step 1 of 5

Name

First

Last

Agency or Organization

Data accessibility

Can you use filters or similar tools in a dashboard, viewer, or other interactive environment to look at pieces of, rather than the whole, dataset? *

Explore adding interactivity. Giving users the opportunity to view and select parts of a large dataset is a helpful practice.

Is the data available for download in rows and columns (e.g., .csv, .xlsx)? *

Data products should provide access to the underlying data — .csv and .xlsx formats are analysis-ready and useful to a broad user base.

Is API access to the data available and clearly indicated on the data product or landing page? *

Consider an API. Though this tool can challening to build, it adds tremendous value for users accessing data programatically.

Is an FTP site available to support bulk downloads and clearly indicated on the data product or landing page? *

Consider an FTP site. Its familiar file structure allows non-technical useres to download lots of data at once.

Is there a landing page with information about the data product that is clearly linked from the data product? *

Make sure there is a central hub for the data product and that everything is clearly labeled and linked. Collecting key information in one easy-to-navigate spot helps the user.

Is the writing on the data landing page "clear, simple, meaningful, and jargon-free?" *

To comply with the Plain Writing Act of 2010, review how the data product is presented. Attempt to remove jargon and aim to educate the broadest possible audience about your data product.

Timeliness of data

Is the amount of historical data available explained? *

If a user sees one year or multiple years of data, they might wonder why those year/s in particular and not others. Add a few sentences to provide the answer.

Is the appropriateness of comparing data over time discussed? Check the box if not applicable (e.g., if the data represent a snapshot in time, if the comparison is provided for the user, or if comparability is otherwise not a concern). *

Seeing whether something has increased or decreased is a common way for users to learn about an issue. Provide comparisons or possibly warnings about constructing multi-year analyses.

Is the update or publication cadence posted? *

The user will want to know how often to expect fresh data. How frequently will the data update?

Is the publication date or date of last update posted? *

The user will want to know how up-to-date the data product is. Be specific in separating “page updates” from “data updates.”

Helpfulness to the user

Are variable definitions provided via codebook, data dictionary, glossary, or key terms list? *

Users need unambiguous explanations of variables to understand the data provided. The best data products specify the data format for each variable and do not data combine concepts (e.g., a value and a year-over-year change).

Are variable definitions clearly linked or displayed? *

Variable definitions are not helpful if they can’t be found. Putting them in multiple places (redundancy) can help, like linked adjacent to a data viewer as well as in the data download itself.

Does the data product address uncertainty through statistical methods or discussion? *

Every data product, statistical or otherwise, has limitations, uncertainty, and methodological decisions that need explaining.

Is a point of contact provided? *

The user should be offered a way to get questions answered should the provided documentation prove insufficient.

Are popular or useful cross-sections, takeaways, or views of the data provided? *

The publisher has the best sense for highlights or insights from the data product — signaling this to the user is a helpful jumping-off point for exploration.

Are there accompanying reports that include relevant data tables? *

Provide a summary report, blog post, or newsletter — this type of data journalism or curation can be helpful in answering research questions.

Is a suggested citation provided? *

Ready-made citations help users understand which program within an agency produced the data product, and reduce the burden on the data user of assembling a citation.

Interoperability

Is the methodology provided? *

Methodology documentation explaining how the data were collected and transformed is a vital piece of metadata that should always be provided.

Is the methodology clearly linked or displayed? *

Having methodology documentation is not helpful if it can’t be found. Putting it in multiple places (redundancy) can help, like linked adjacent to a data viewer as well as in the data download itself.

Is other metadata (i.e., anything that helps with interpretation and interoperability of the data, like an FAQ or crosswalk table) provided and clearly linked or displayed? *

There are always opportunities to provide additional context or detail to support machine readability or the user’s understanding of the data product.

Are variable names both short and meaningful (to facilitate quick scanning and human readability)? *

Done right, condensing variable names while retaining meaning makes it possible for the user to understand the data in situ without heavily depending on a data dictionary. The best data products are also consistent with respect to spacing and capitalization of variable names, to further support human readability.

Are variable names free of special characters (anything besides underscores; e.g., $, -, *) to facilitate machine readability? *

Computer code will often use special characters — by including these in variable names, it becomes more difficult to programmatically work with data.

Total

$0.00

Back to Federal Data Excellence home