An Open Data Menu for Arkansas
Last Thursday the Data Coalition was honored to meet with the Arkansas Open Data and Transparency Task Force. The Task Force is a one-of-a-kind body, established by law, that includes state legislators, representatives of the Governor and Attorney General, and the leadership of most of the state’s largest agencies. Its job: by the end of this year, recommend an open data law to the Arkansas legislature.
The Task Force invited the Data Coalition to present a “menu” of open data options. With such an open-ended mandate, where should the Task Force focus its reform recommendation? What are the benefits of standardizing and publishing public-sector spending information? What if the Task Force focuses on regulatory reporting instead? How does health care fit in? We recruited government leaders and Coalition members to explain the benefits other states, the federal government, and foreign countries have derived from open data in each substantive area.
We told the Task Force that open data isn’t just good for public accountability. It also powers analytics for internal management and – done right – reduces compliance costs by automating reporting tasks that used to be manual.
Here are the presentations we shared with the Task Force on Thursday – and what we learned ourselves.
Open Data in Spending: Deliver Transparency, Stop Fraud
The federal government, and many states and localities, are adopting consistent data standards for information about their finances, transactions, grants, and contracts – and publishing that information as open data. When spending information is expressed as open data instead of old-fashioned documents, citizens can use it to hold politicians accountable; managers and inspectors general can visualize and analyze spending more easily; and required reports by grantees and contractors can be automated.
The federal DATA Act, whose passage our Coalition celebrated in 2014, requires the executive branch to transform all federal spending information into open data. The U.S. Treasury Department and the White House Office of Management and Budget are hard at work creating a government-wide data structure to bring the federal government’s disparate types of spending information together as one searchable data set. Federal agencies must report their spending using these standards by May 2017. Meanwhile, the law requires OMB to run a pilot program that determines whether the same standards can help federal grantees and contractors automate their reports and comply more cheaply.
But many state and local governments are far ahead. Seth Unger, senior policy advisor to Ohio Treasurer Josh Mandel, showed the Arkansas task force that Ohio’s Online Checkbook allows citizens to navigate through all spending by the Ohio government and its agencies – from statewide categories all the way down to each individual transaction. By publishing spending information as standardized, open data, the Online Checkbook improved the state’s grade in the Public Interest Research Group’s ranking of state spending transparency efforts from D minus to the first-ever A plus.
Mr. Unger told the task force that Ohio’s Online Checkbook has allowed users – both inside and outside the state government – to quickly notice “outliers” and strange patterns and take corrective action.
(Ohio had an easier challenge than some other governments because it had already standardized its spending information. Ohio adopted a single, consolidated statewide financial system in 2007 – a far cry from the thousands of incompatible financial systems in use by federal agencies. But even with a single financial system in place, Treasurer Mandel and his team had to improve data standardization to make sure the data set on Ohio’s Online Checkbook was meaningful and accurate.)
Treasurer Mandel has taken Ohio spending transparency to the next level by inviting Ohio’s local governments to voluntarily publish their spending as open data on Ohio’s Online Checkbook. Nearly 100 cities and towns, 75 townships, 15 counties, 70 school districts, and 15 other entities have done so.
Even in advance of the DATA Act’s mandate, many federal agencies are standardizing internal data sets to empower analytics to suss out fraud. Dave Williams, former inspector general of the U.S. Postal Service, told the task force that he deployed multiple analytic tools against the Postal Service’s spending data. RADAR assigns a fraud risk score to each of the Postal Service’s hundreds of thousands of contracts, allowing investigators to focus on those most likely to involve fraud. TripWire sends an alert when specific situations – such as a proportionally large contract modification or a sudden increase in an employee’s health claims – occur.
“When information is standardized [as open data],” said Mr. Williams, analytics “will be immediate.” But without data standards, every analytics project requires expensive, purpose-built translations.
Open Data in Regulatory Reporting: Cheaper for Business, Better for Agencies
When regulatory agencies choose to collect reports using standardized, open formats, instead of unstructured documents, the information becomes easier for the agencies to analyze – and can be reported in an automated fashion.
Pramodh Vittal of DataTracks told the task force that the UK tax authority, HMRC, simplified reporting for millions of companies by adopting the inline XBRL (“iXBRL”) open data format. This format is both human-readable and machine-readable, which means it can fulfill document-based reporting requirements while also empowering analytics software. Mr. Vittal’s presentation showed how software helps companies easily report their financial information in iXBRL.
(HMRC is working side-by-side with Companies House, which regulates all companies in much the same way as most U.S. states’ secretaries of state, to collect tax filings in the same format as Companies House’ corporate filings. This 2012 video from Companies House explains how it works.)
A similar story comes from the Washington state Utilities and Transportation Commission (UTC), which is running a pilot project to adopt an open data format for the financial information that utilities must submit. Although the Washington UTC could not join our presentation on Thursday, an eight-minute video produced by Coalition member RR Donnelley explains how the open data format improves accuracy and reduces costs.
Open data is coming to federal regulatory reporting on a grand scale. Last year, open data supporters in Congress introduced the Financial Transparency Act (H.R. 2477), which will require all eight major federal financial regulatory agencies to adopt consistent data standards for all the information they collect from public companies, banks, and financial firms. Open data in financial regulatory reporting will bring transparency to investors, enable analytics to anticipate Enron-style frauds, and allow regulated companies to automate compliance.
Last week the Data Coalition’s Financial Data Summit brought together over 300 supporters of this transformation. Financial Data Summit speeches, presentations, and media coverage are all available on the Coalition’s event site.
Health Care, Parks and Recreation, and Anywhere Else: Standardizing and Publishing Makes Life Easier
Mary Kay McDaniel of Cognosante told the task force that data standards for health care information can drive new functionality, like price comparisons and quality ratings. Her presentation introduced the main standards in use today, explored the reasons why data standardization is challenging in health care, and offered examples of standards-driven apps. She encouraged the task force to send a representative to national health care data gatherings like the Health Datapalooza.
Although campsite-reservation startup Hipcamp could not send a representative, we explained to the task force that if the Arkansas Department of Parks and Tourism were to publish campsite availability as standardized data, Hipcamp and other startups could use it to enable consumers to make electronic reservations. The national park system and California’s state parks currently offer standardized campsite availability data. But Hipcamp is still asking all other states to take that same step.
What do health care information and campsite availability have in common? Both become more useful to consumers, professionals, and public servants if standardized and published. In fact, as we repeated to the task force, standardization and publication solve problems in pretty much every area of government data.
Open Data Across Departments, Divisions, and Disciplines
The final phase of our presentation to the task force introduced four open data platforms that aren’t limited to a specific type of information. Tableau, Socrata, Esri, and OpenGov all allow users to organize and visualize information and make connections across multiple data sets – connections that would be impossible if those data sets weren’t standardized and published.
Open data reforms must be purpose-built for complex areas of government information, like spending, regulatory reporting, health care, and recreation. For instance, the federal DATA Act had to call out specific reports and systems and provide a strong mandate for data standards within spending.
But Tableau, Socrata, Esri, and OpenGov all showed the power of data sets sourced from different areas of government information, combined using common standards, and then deployed for the public good.
Anthony Young of Tableau used the Florida Department of Transportation’s auto crash dashboard to show how standardized data can deliver unprecedented insights for government leaders. Tableau’s platform allows Florida policymakers to categorize, visualize, and understand auto accidents statewide and by county. The data standards aren’t perfect: for example, when the department changed its reporting methodology for distracted driving in January 2011, trendlines jumped. But the standards are sufficient to reveal insights that could never come from old-fashioned document-based reports.
Chris Rodriguez of Socrata used the state of Iowa as a use case for connecting data sets from different disciplines. Iowa’s original open data site required users to know what agency published the data they were looking for. But the new data.Iowa.gov portal allows users to run Google-style keyword searches or browse by topic area. The portal can lead users to interactive open data off all sorts, from unemployment insurance statistics to granular spending information. And it connects automatically to the federal open data portal, Data.gov – which means Iowa’s data sets can be combined with other states’.
Mr. Rodriguez shared a surprising observation from Iowa: most users of the state’s data portal are state employees working inside government – not citizens following their government. These state employees are finding data.Iowa.gov a more versatile than the internal systems from which the data sets are sourced. Socrata’s observation matches one of the most popular goals of the open data movement: governments should “eat their own dog food,” basing management decisions on the same information they publish as open data.
Matt Bullock of Esri introduced the state of Michigan’s open data portal, which Esri maintains. Esri’s specialty is geospatial information. Since the “vast majority of [government] data has a geospatial component,” Michigan’s portal offers many examples of combining different data sets on one map. Michigan’s portal also allows embeddability: every data set includes electronic codes that allow it to be featured in new websites, automatically linking back to the original data set for automatic updating whenever the source changes.
Mike Dougherty of OpenGov explored the open data network effect: when multiple governments publish their information using the same standards, comparisons among them become possible. OpenGov’s 900 clients, including several Arkansas towns, have all embraced the same data structure for accounting, transactional, and performance data – which means their budgets and finances can all be compared against their peers’.
OpenGov’s client list also includes Treasurer Mandel’s success story in Ohio.
What’s Next for Arkansas?
The Arkansas Open Data and Transparency Task Force is going to deliver a report to Governor Asa Hutchinson and both chambers of the state legislature by December 31, 2016. The report will recommend legislative language for an open data law.
Open data policy reforms are a challenge because the benefits of open data are as broad as government information itself:
- Open data can deliver better accountability, as Ohio’s experiences show. Ohio’s online checkbook makes outliers more visible than they ever could have been if trapped in old-fashioned document-based records.
- Open data can deliver better management. Iowa’s state employees are the most enthusiastic users of the state’s open data portal – even though they presumably also have access to the internal systems from which the portal sources its data sets. Clearly, when information is standardized and published, it becomes more useful to managers within government as well as citizens outside government.
- Open data can deliver automated compliance. UK companies can report to Companies House and HMRC simultaneously because those two agencies adopted the same open data format for financial information. That’s cheaper than filing two separate reports.
The Open Data and Transparency Task Force’s job is challenging – not because it’s difficult to derive benefits from open data but because it’s so easy! There are so many possibilities for better accountability, better management, and automated compliance throughout the state that nine more months won’t be enough time to explore them all. There is too much on the menu.
But we hope Thursday’s presentations provided a good start.