How Federal Agencies Can Better Manage Data
Federal agencies can tackle these challenges through the traditional trifecta of people, processes and technology, Saliter says.
Anecdotally, he says, the public sector knows that it has “not been on the cutting edge of going after and attracting” personnel with skill sets in emerging technologies. Agencies need to do more to attract such workers and compete more with the private sector, according to Saliter.
In terms of technology, agencies need tools that are capable of performing data integration and handling higher volumes of data. “We as an industry need to be able to handle much more volume than exists today,” Saliter says.
Agencies that are at capacity in terms of computing performance or data storage need to rethink their IT architectures, he adds. IT leaders also need to invest in technologies that enable real-time analysis and decision-making.
Historically, agencies relied on transitional data warehouses or data lakes to dump that data into a repository. Agencies then used reports or business intelligence to extract data and get value and insights.
“There is a time and place for that,” Saliter says, but in general, in “that architecture is not going to be able to keep up with the volume and rapid decision-making that is expected going forward.”
Now, agencies need to have solutions that can ingest real-time performance indicators, known as metrics and traces, that are provided by applications. Saliter also says agencies should invest in streaming technology to handle data in motion and be able to analyze data in near real time. Another path forward is to invest in AI and machine learning to detect anomalies and perform projections.
Agencies need top-level management support of these initiatives, Saliter says. “They need to make sure this becomes a cultural change. They need to bring in diversity in thinking and diversity in skills.”
Some of that will come by agencies hiring younger workers and data scientists out of college who can bring fresh thinking, Saliter says.
READ MORE: Find out how feds are learning new data science skills.
Federal Data Strategy Can Help Evolve Agency Efforts
The federal government is in the midst of implementing a 20-point action plan to start making the Federal Data Strategy a reality. The strategy “provides a common set of data principles and best practices in implementing data innovations that drive more value for the public,” according to the Office of Management and Budget.
The Chief Data Officers Council has also provided a platform for agency CDOs to collaborate and share best practices. According to a survey of federal CDOs released in August, more than half of the CDOs surveyed reported improvements in data quality (64 percent), assessment of staff capabilities and needs (57 percent), migration to cloud-based services (57 percent), and availability of metadata (54 percent).
A combination of more empowered roles for CDOs and regulations or policy can have an impact on agencies’ ability to handle large data volumes, Saliter says.
The Federal Data Strategy itself is “going to stimulate and put fuel on the fire to move toward a more analytics- and data-driven approach,” Saliter says. “I would echo that policy is critical to this happening even faster and further.”
MORE FROM FEDTECH: What is predictive analytics and how can it help agencies?
