Government Thought Leaders talk Big Data
Published: October 24, 2018
Government leaders came together at an event hosted by ATARC to discuss big data methods and technologies being used by their agencies today.
The Federal Data & Analytics Summit hosted by the Advanced Technology Academic Research Center on October 23rd brought together a variety of government speakers and panelists. Keynote topics such as the new federal data strategy and the direction the government should take in artificial intelligence were explored. Candid remarks were made by government panelists on data innovation and disruptive technologies and questions from the audience, most of whom represented other federal agencies, were riveting. The general agreement among federal agencies is that a more data-driven government is needed for key decision-making purposes. However, the event also helped shed light on the challenges and different approaches in how to achieve such a state in government.
Government speakers agreed that they have a long way to go in achieving full use of the massive amounts of data that has been collected over the years. However, several success stories emerged to prove just how agencies have tackled some of their big data dreams.
Marina Fox from GSA stated that the agency’s new Solicitation Review Tool (SRT) will be in full production on a cloud platform in 2019. Fox shared that the SRT needed a variety of disruptive tools such as AI, ML and NLP to build the predictive engine which will detect non-compliance with Section 508 law in the solicitations issued by the federal government. Lesson learned along the way included being vocal about changing an agency’s planned roadmap to allow for flexibility in technology to achieve innovative projects. Likewise, Col. John Scott, a Data Manager at DHA, stated that despite what is perceived in Congressional hearings, DOD has advanced in its collaboration with the VA. For instance, DHA has established an electronic file cabinet for health records that are human-readable. Moreover, DHA has had much success moving many of its records to VA services members.
Nonetheless, government leaders expressed frustration over issues such as interoperability. Col. Scott stated that the biggest challenge in the DOD/VA EHR operation is getting all the data into an integrated data platform. If done the smart way, Scott said that the DOD/VA will have an advanced and unmatched healthcare capability. Leonal Garciga, CTO at the Joint Improvised Threat Defeat Organization, agreed that scaling across organizations is difficult. However, Garciga stated that agencies must change the method of building out from the bottom of the stack and use the technologies that will help build across stacks. Nevertheless, he concluded that despite interoperability issues and legacy infrastructure remaining, the government has come a long way in integrating their data from five years ago.
When asked how the government panelists are balancing governance/security of data with flexibility to allow for innovation, all participants agreed this is the most challenging area and that a firm solution is not yet in sight. Col. Scott stated that his team is trying to work with a commercial product to standardize some data and identify the data that can be allowed for innovation and has hit various limitations in terms of governance to achieve this. Similarly, William Spencer, IT Category Program Manager at GSA, stated that he faced the same problem when trying to build a data assessment model and needed answers from various managers across organizations. For him, the key to success in obtaining responses for a functional assessment was acquiring executive-level backing. At the onset, his team set specific objectives with the questions aligned to answer the objectives based on confirmation from his executive line. Garciga stated that the best way to deal with governance is by first, “opening up the aperture.” His viewpoint is to open up the data and give people, particularly millennials, the flexibility to identify innovative data analytics techniques and then place the governance over data afterwards. An agency’s security ecosystem must be designed to allow for build outs of the data, said Garciga.
Col. John Scott shared some current analytical projects at the DHA. The first project has to do with the agency’s disability system. DHA is trying to study service members that left due to medical disability and apply predictive models to determine how those service members got to that point and what can be done to prevent the situation ahead of time. The second project includes working with VA to identify the veterans that are at great risk of having a difficult time after the military by studying the last few years of service and again, setting up predictive models to help transitions to the VA.