The topics of big data and predictive analytics come with big promises of delivering new and greater insights for government agencies. While the capabilities are real, results won’t happen overnight – there is a methodology to implementing big data analytics. Achieving desired outcomes requires a strategic approach.
In Wednesday’s GovLoop training, “Delivering on the Promise of Predictive Analytics: Batch & Real Time Analysis,” experts examined how batch and real time analytics impact the outcome of predictive analytics by identifying the role each plays in the analytics process and examining real world examples of how government is benefiting from these tools today. They also discussed strategies for implementing predictive analytics within government agencies, that can help remove complexity and expedite time to value. The session was part of GovLoop’s Government Innovators Virtual Tech Day.
Dennis Vega, Managing Director for Planning, Performance, and Systems, Office of U.S. Foreign Assistance Resources U.S. Department of State, discussed how using data to help make decisions for foreign assistance is especially critical. “We’re moving past the recognition we need to use data, to actually using the data,” Vega explained. Vega quoted the 2015 Quadrennial Diplomacy and Development Review, which stated, “Making progress on U.S. policy priorities will require a data-driven, evidence-based approach. The amount, availability, and variability of data is expanding exponentially, and it has the potential to inform us about a range of issues, including conflict dynamics, state fragility, corruption, popular opinion, and climate change.”
Vega said the Department of State uses a broad set of sources for data when using it to make decisions, including country performance data, financial data, foreign policy priorities, program performance data, and more.
Vega pointed to one of the Department’s sources of data, ForeignAssistance.gov, a public data portal that houses and visualizes U.S. foreign assistance data. It contains tens of thousands of transactions published on a quarterly basis and helps improves internal operations and decision-making within the Department of State and the 20+ other U.S. agencies that manage foreign assistance. “Using the data there helps recipient governments manage incoming aid and make budgeting and resource allocation decisions,” Vega said.
While data helps the Department make better decisions, it’s not without its challenges, Vega said. “Not all decisions can be data-driven,” he said. “Policy and politics plays a role. Additionally, real-time information is not always available – or accurate, and proving the value of data quality efforts to internal stakeholders can be challenging.”
But it’s getting easier. “There’s a progression of a culture of using data in everyday life in a way there hasn’t been before,” Vega said. “This makes it easier to help people understand the value of making data-driven decisions.”
Next up was Clint Green, Director of Advanced Analytic Strategy and Development ViON, who discussed the meaning and definition of analytics, and how understanding how your organization makes decisions is key to knowing where to apply advanced analytics.
There are actually four kinds of data analytics, Green explained: description, diagnostic, predictive, and prescriptive. They answer, respectively, these questions: What happened? Why did it happen? What will happen? And how can we make it happen?
These four types of data analytics all matter. You have to understand what happened, to understand why it happened, and then understand what will change or happen in the future. Trying to understand everything with the angle of prescriptive analytics is tempting, Green said, but he cautioned folks to move slowly in analytics. For a data analytics solution to improve accuracy over time, it must build on the previous results and learn from that. The results of previous analysis, become part of the data pool for future analytics and over time, a cycle develops where the data analytics solution learns from itself.
It’s hard to say with certainty that will happen a month from now or even a year from now. But agencies can use the data they have to better predict what may happen in the future. That’s the value of predictive analytics. It helps agencies plan accordingly in a variety of areas, from budgeting, to future customer service demands, or internal HR planning. But knowing what may happen does an agency no good if it’s unprepared. That’s where prescriptive analytics comes into play.
Green closed with a quote fromWilliam Bruce Cameron: “Not everything that counts can be counted, and not everything that can be counted counts.”
