The Open Government Directive: Make a Plan for Your Data

The Obama administration is close to unveiling a new Open Government Directive to push agencies to become more transparent, accountable, and collaborative. The directive builds upon the memorandum on transparency and open government released by the administration in January, and was announced by US CTO Aneesh Chopra during the Gov 2.0 Summit. Since the announcement, there has been a lot of discussion focused on the expected component of the directive that pushes federal agencies to make data available in machine readable formats. This is a huge opportunity that will help make it even easier for third parties to create valuable applications to aggregate, organize, and visualize government data. Our experience building DataMasher as part of the Apps for America II contest was extremely positive, but one of the biggest challenges for us was transforming the data on Data.gov into a useable format. See our recent post with some more thoughts on this. Pushing federal agencies to do more of the legwork up front will help third parties focus on creating value with the data rather than just burning hours to transorm it.

However, some are concerned about how far agencies will have to go to transform legacy data, especially when it is in inaccessible formats such as PDF. This concern is understandable, but it shouldn't delay efforts on the part of federal agencies to take action. In planning their response to the directive, we'd recommend that agencies go through a series of steps:

• Do an inventory - valuable data is often dispersed throughout the organization, sometimes "owned" by people with little incentive to share it (or are just unaware of the need).
• Identify and prioritize - evaluate the utility of different data sets to determine which has the highest potential value to the public and assess the difficulty to transform it or make it available in an accessible format. There might be different approaches for opening up different data, including APIs, XML feeds, RDF format, or in some cases CSV.
• Develop a plan - based on the priorities, develop an ongoing plan for making the data available, starting with the easiest and most valuable data.
• Just do it - whether submitting existing data to Data.gov, planning an API, or conducting pilot projects to transform and make data available. Start with the low hanging fruit immediately to show progress, and work in parallel to make other data available.
• Highlight progress and build upon small successes. This will take time and effort, organizations should promote their progress so far and communicate to the public the steps they are taking.

Ideally the organizations' plans to make data available in machine readable formats should be part of a larger plan (that goes beyond data) for being more transparent, participatory, and more accountable. But that is a topic for another day, and another post. Stay tuned.