Reviewing the adoption of Sunlight’s Open Data Policy Guidelines

city skyline in data drawing
(Image credit: iStockPhoto)

While cities around the nation are “geeking” their way to excellence, we on the Sunlight’s Local Team track the passage of open data policies. But how effective are the policies these cities are passing? We decided to find out.

When we find newly enacted policies we enter them into a database of open data policies and add them to our open data policies at work page. To compare policies to each other, we comment in the database on which of our 31 Open Data Policy Guidelines are referenced and which recommendations are not included within the language of policies. Using this database as a foundation, we were curious as which of our guidelines were being followed more often than others. So, we dug into the numbers.

To establish “Guideline Adoption Rate,” we assigned weights to different kinds of content in the spreadsheet. Cells that contain “close fit” were assigned 1 point, “N/A” were assigned 0, and “the opposite” guidelines were assigned -1. As 66 cities are included in the spreadsheet, we subtracted the “N/A”, “close fits” and “opposite,” then got the number of policies that meet the guidelines, and assigned 2 points each. Finally, we calculated the sum and divided it by 132(66*2), which is the maximum points possible for each guideline. Here are some major takeaways from this research as of August 2016.

To view the full table in a new window, click here.

Most adopted open data policy guidelines

#1 Create a central location devoted to data publication and policy

85.6 percent guideline adoption rate

We often see open data portals precede policy adoption. This guideline enhances the longevity of open data portals. The following policies do not have language promoting this guideline:

  • Charlotte, N.C.
  • State of Delaware
  • Lexington, Ky.
  • Memphis, Tenn.
  • State of New Hampshire
  • Portland, Ore.
  • Seattle, Wash.
  • State of Texas

#2 Mandate data formats for maximal technical access

83.3 percent guideline adoption rate

A core principle for the open data movement is the ability for data to be “analyzable” by computers in common formats. Many policies offer this tenet within their policy language. However, the following policies do not have language promoting this guideline:

  • Connecticut
  • Providence, R.I.
  • Memphis, Tenn.
  • State of Rhode Island
  • Suffolk County, N.Y.

#3 Appropriately safeguard sensitive information

82.5 percent guideline adoption rate

This is an indicator of what the top concern is for local officials. As stewards for public information, governments take the responsibility of information security a high priority. We offer resources on this topic, and actively follow good examples of governments safely releasing microdata. The following policies do not have language promoting this guideline:

  • Amherst, N.Y.
  • State of Delaware
  • Detroit, Mich.
  • Memphis, Tenn.
  • Minneapolis, Minn.
  • State of New Hampshire
  • Providence, R.I.
  • Raleigh, N.C.
  • State of Rhode Island
  • Williamsville, N.Y.

Least adopted open data policy guidelines

#1 Require digitization and distribution of archival materials

0.8 percent guideline adoption rate

Howard County, Md., has this guideline within its policy. Storing bits can be less costly than storing paper. While the upfront cost of converting paper records may seem high, compared to the cost of records storage and staff, records conversion can save money, increase efficiency and reduce the cost of storing records over time.

#2 Require publishing data creation processes

1.5 percent guideline adoption rate

Seattle is the only policy that addresses this guideline. The professional equivalent of “showing your work” offers an opportunity for widespread review to validate the methodology on which datasets or policy decisions are based. If the goal is “data-driven decisions,” then citizens must know how the data was compiled, filtered and transformed.

#3 Mandate the use of unique identifiers

1.5 percent guideline adoption rate

Kudos to Seattle, which again is the only policy that calls for unique identifiers. Implementing unique identifiers can be easy, assist citizens’ ability to analyze records between datasets and relates well with the highly adopted guideline on safeguarding sensitive information despite the Mosaic Effect.

The Open Data Policy Guidelines and the Open Data Policy database are actively updated resources. This review help us understand not only where open data policies don’t address our best practices, but where Sunlight can focus and update the guidelines in the next version to usher governments towards well rounded policies that improve all facets of openness.