Data-driven Decision Making at Work in the Public Sector
Governments around the world are adopting DDD in their daily operations, not only to react to citizen demands and concerns, but also to proactively anticipate an issue before it develops into a crisis. Here are some examples of advanced data-driven government solutions happening right now within public agencies to evolve the role of government to be more citizen-centric.
The Hong Kong government's Efficiency Unit acts as single point of contact for many government departments to handle citizen complaints and suggestions. Each year, the unit receives 2.65 million calls and 98,000 emails. "We gather substantial volumes of data. The next step is to make sense of the data," says the Unit's Assistant Director, W. F. Yuk. The office partnered with a text mining firm to build a complaint intelligence system to uncover root causes. Report generation is shortened from one week to one click, and the responsible department is immediately informed of issues.
Transportation is an area where a broad range of challenges can be tackled by DDD, such as traffic flow models to reduce congestion, setting the price of tolls, and finding optimal routes to destinations. For example, West Virginia's Department of Transportation continuously tracks traffic at 2,500 spots around the state to understand a variety of components such as average daily traffic, vehicle type information, intersection turning movement information, and annual vehicle miles. The information is used to plan infrastructure enhancements and prioritize construction projects.
Utilities and Energy
Energy production must be balanced according to demand. Some products like electricity need to be consumed as they are produced; otherwise surplus product effectively goes up in smoke. Meanwhile, other products such as gas can be stored, but the cost of storage can be high. On the other hand, a disruption in supply can cause huge economic damage.
Eastern Denmark previously had 16 partners to balance electricity supply on a daily basis in order to anticipate the right amount of power consumption and production needed. After partnering with Copenhagen Energy, which emphasized the use of DDD, consumption can be predicted on an hourly basis to minimize production waste.
Another critical application for DDD in utilities involves encouraging citizens to manage their own consumption and be proactively notified when they reach their normal consumption thresholds. This can happen when an electrical appliance is left running while people go on vacation. Through the analysis of citizen utility consumption patterns, DDD can proactively identify early signs of overconsumption. Notifications can then be sent to citizens through SMS and mediation action can be taken to avoid bill-shock.
A country's continued development and international competitiveness rely on the availability and quality of education across primary, high school, and university levels. Students have different developmental needs and interests across different grades or even in the same grade. Delivering education through a blanket approach does not help every student reach their full potential. DDD can be used to determine the optimal location for new schools and universities, evaluate whether curriculums are meeting the desired goals, and plan teaching staff recruitment based on evolving needs.
The Tennessee Board of Regents is piloting predictive analytics at the university level to match students with appropriate programs, even down to which professors may be best suited for them. The data will also point out how and where universities need to improve in order to graduate more students.
The volume of South Korea's imported goods had doubled over the last decade, but the size of its customs force remained the same. Samples were inspected for illegality based on past illegal importer names, items, and foreign providers. If any of the data points changed, an inspector would likely miss this during inspection. The agency adopted a DDD approach to detect illegal imports intelligently. Now, more than 400 inspectors are using DDD, which apply more than 60 search conditions for risk ranking. Since then, South Korea's detection rate for illegal cargo increased by more than 20 percent.
In California, the Los Angeles Police Department's Real Time Analysis and Critical Response Division uses statistical prediction techniques to create maps with specifically highlighted areas where certain types of crimes are most likely to occur. And in Boston, the Smart Policing Initiative involves longitudinal analyses of violent crime hot spots in the city. As a result, Boston saw a 17 percent reduction in the total number of violent crimes and a 19 percent reduction in the number of robberies during the time studied by researchers in those areas.
Not to be outdone by Boston, New York City Police's Compstat program uses compilation, distribution, and utilization of real-time data to allow police officers to make better informed decisions. Previously predominant measures of success for police commanders were predicated on the number of arrests made after a crime; Compstat helped in changing the measure to crime reduction and the most effective ways for achieving that objective. In just five years the burglary and murder rates had dropped by more than 50 percent. This far outpaced the change in rest of the country. Compstat led to individual departments becoming more empowered for better decision making that can "analyze, reflect, learn, and change based on experience."