Bringing Data Science to Government

April 20, 2016 Amy Deora

I recently joined the team at Civis Analytics as our Director of Government Analytics and I’m excited to bring my background in information technology, analytics, and city planning to help government organizations become more data-driven. For the past ten years, I’ve been implementing large data analytics projects to improve the performance of programs at Federal government clients and I’m looking forward to introducing government agencies to the power of the Civis data science platform and our team of data scientists.

People Science in Government
At Civis, our primary unit of analysis is the person. This means that the software we’ve developed and the insights we’re best suited to provide fit perfectly with local government strategy, since local government is ultimately about serving people. Many local governments are already using data analytics to understand the constituents who reach out to them. Our focus at Civis has often been helping government find the citizens that are not interacting with government, and help give them the information they need to make the best decisions for themselves and their families.

This people-centered approach fits in perfectly with the new drive across government for further adoption of insights from behavioral economics. Behavioral economics involves using small, low-cost changes in program design that can have huge impact, and that don’t involve using the heavier-handed levers of policy, regulation, or economic incentive. For example, a text message reminding college-accepted high school graduates to complete tasks such as completing financial aid forms and taking placement tests led to an 8.6% increase in the number of low-income students successfully enrolling in college, compared to those who received no text reminders.1Participation rates in 401k plans at employers who ask employees to “opt out” of their 401k plan are much higher than those who are asked to “opt in”. These small changes in program design and outreach can have profound changes in people’s lives, at very little cost to the government.

For these types of interventions to succeed, they need to reach the right people, as quickly and efficiently as possible. This is where Civis comes in. Predictive modeling has been used for many years in government to identify patterns that identify fraud, risk, or lawbreaking. However, the next step is using analytics to proactively identify citizens who simply need additional information, outreach, or tools to reach their goals or interact more efficiently with government.

Our predictive models help identify which people organizations should be reaching out to, including government organizations. For example, we have helped a state agency in their efforts to conduct outreach to families who could afford to, but are not using tax-advantaged methods to save for college, and helped cities and states identify the uninsured after the advent of the Affordable Care Act. Similar methods can be used for many other applications, such as ensuring that veterans are aware of their benefits, identifying students who might be at risk of dropping out of college, or identifying taxpayers who need to file a particular form but may not realize it. Combined with our state-of-the-art message testing and experimental design services, we can ensure that the the right message gets to the right people.

Our path forward
Although the speed of change can be slower in the public sector than it is in the private one, the public sector has seen a recent increase in interest in using more data-driven decision-making. Government has also seen a recent influx of talented data analysts with an interest in data science and exposure to the types of models and software tools we, at Civis, use every day. Agencies are getting savvier about data analytics all the time – they don’t just want to a see a report with results, they want to see what makes a model tick under the hood.

Government is more interested than ever in working with the data themselves, and having “self-service” access to analytics products – which is just the way we like to work at Civis. For this reason, we share our data science tools, including the Civis data science platform, with our clients. We are looking forward to continuing to provide the data science know-how and software tools to help government agencies at all levels really make an impact.

The post Bringing Data Science to Government appeared first on Civis Analytics.

Previous Article
It was #AlwaysTrump
It was #AlwaysTrump

Tuesday night all but officially certified Donald Trump as the Republican nominee for president. While Cruz...

Next Article
Ask America Anything Results: How the US Feels about Gun Control & Domestic Violence
Ask America Anything Results: How the US Feels about Gun Control & Domestic Violence

In December, we invited nonprofits to Ask America Anything by giving away two questions on our weekly natio...