Open Government, Open Data and Some Lessons from Arkansas

I was reading Brian Timoney’s excellent post “Open Government is a Slammed Door at the BLM,” when I encountered this line:

Top-down, large scale spatial-data-integration-by-directive simply doesn’t have the track record of success to justify the costs…

This resonates well with observations that I have made during the course of my recent work with the Arkansas GIO office. The immediate focus of that work was to help AGIO assess the current state of their GeoStor platform and gather data about current and emerging capabilities in order to support decision-making about whether to migrate to a hosted (cloud) environment. In the course of doing that work, I learned about the collaborative approach AGIO has taken with the cities and counties of Arkansas, which also informs some of the functional requirements of GeoStor.

AGIO, among many other functions, offers state-wide, seamless data sets for streets, parcels and addresses. These data sets are assembled from county/city data and allow GeoStor to feature a highly accurate geocoding service, which serves as a resource to support requirements such as streamlined sales tax and others. AGIO feels that commercial geocoding engines do not provide enough accuracy to support these types of requirements so they use detailed data from local jurisdictions to build one for the State.

What I came to understand is that AGIO integrates these data sets in the absence of any top-down requirement of the kind Brian mentions in his post. In general, they take the data as it comes (format, schema, etc.) and use automated/manual processes to transform the data and build/update the data sets. Using this “light touch” approach, AGIO receives data from all 75 counties in the State.

This arrangement works because AGIO views GeoStor as a state-wide asset that must be able to provide value back to the localities as well as to individual citizens. Local jurisdictions can store/host their data on GeoStor. They can also make use of the platform to host map viewers and applications. For instance, the Saline County Assessor’s Viewer is hosted on the GeoStor platform. So that geocoder can be used by the localities for their own needs without needing to build and maintain their own. Another example is the GeoStor search tool, which can be used to expose any asset registered with GeoStor (local or linked). This search engine becomes a single point of entry that counties can use to expose their own products and services.

And, again, all of this works without a top-down approach of “integration-by-directive,” as Brian puts it. When I asked Learon Dalby about this approach, he cited minimal impact on counties and value-add, among others, as two important core values that inform the work AGIO does. This approach requires a lot of leg work. Staff visit the counties and AGIO seems to be remarkably in touch with the needs at that level. Basically, all participants get something of value out of the arrangement and the citizens of Arkansas benefit from the collaboration.

This approach stands in stark contrast to much-less-successful efforts I have witnessed at the Federal level. Through my participation in various working groups over the years I have seen numerous top-down efforts fail. Commonly, this is because they offer a data standard that legislates both content and structure. Oftentimes, these standards to do not mesh well with data that is already in place at lower echelons so additional collection/validation/conversion procedures are required, leading to increased complexity and “discussions” over who will fund/execute that additional complexity. Since complexity usually translates into dollars, the stream of vendors and integrators with solutions to mitigate that complexity soon follows.

Another situation I have commonly observed in such efforts, especially those that attempt to get data from the state level up to the Federal level, is a lack of perceived value by the state-level data providers. Often, they feel they do not see value flow back down. They see high-quality data go in and little, if anything, come back out. This is often obscured by nebulous policy or security requirements but almost universally leads to frustration.

Which brings me back to the Arkansas model. The minimization of top-down requirements and the focus on providing clear value (which doesn’t always need to be realized in dollars) to all participants are key reasons why it works. A straw-man that is often tossed out in working groups is “Will it scale?” That’s a non-factor here. With 75 counties participating, it is already operating at scale. Here, I am not referring to data. Scaling data is easy. Scaling relationships and participation is harder. Processes that impose little overhead scale better. AGIO succeeds by knowing their mission and pushing down as few requirements and as much value as possible.

If the Federal government (including the defense and homeland security sectors) ever gets serious about “open government,” “open data,” data sharing and other such efforts, it might be worth a trip to Arkansas.