Supporting public policy with predictive analytics
When we talk about quality of education, we usually talk about what happens inside a school: how well trained the teachers are, how current the materials are, what kind of access to technology the students have. But many experts, including Dr. Patricio Rodriguez at the Center for Investigation of Advanced Education (CIAE) at the University of Chile, believe the quality of education in a given school is directly connected to the socioeconomic and demographical factors outside the school.
To improve the educational system in the greater metropolitan Santiago area, the Chilean Ministry of Education asked Dr. Rodriguez to help decision makers better understand the context in which the city’s schools are operating.
They wanted to use data to reveal previously hidden relationships between socioeconomic factors like urban planning, crime, and transportation to the quality of K–12 education in metropolitan Santiago. They wanted to know where to build the next school, how to improve transportation in remote areas, and how to invest in the region to improve access to school and prevent dropouts.
The Ministry wanted to see what answers the data might provide—where school-aged children live and how they get to school—but it didn’t have a way to combine the data for meaningful analysis or present it in a sensible way to decision makers. This is why it looked to the CIAE at the University of Chile to help analyze historical information to inform budgetary and policy decisions.
Using Microsoft Azure services, including Azure Machine Learning and Microsoft Power BI, Dr. Rodriguez and his team built explanatory and predictive models to see what they could learn from the data. After analyzing the geography of school-aged children and the location of the city’s schools, for example, they found that 32 percent of students do not have a school within a 10-minute walk from their house. Additionally, the data showed the majority of schools deemed higher performing schools—those in which at least 30 percent of students achieve an adequate proficiency level in national standardized tests—were clustered toward the center of the city where socioeconomics were higher, leaving the students on the periphery with fewer options.
By posing questions of the data and testing its hypotheses in Azure Machine Learning, the Ministry of Education in Santiago can now see for the first time the connections between specific public policy decisions and educational phenomena, such as school attendance and drop-out rates. The Ministry can then easily visualize the findings for stakeholders through Power BI. Having these insights allows the government to develop specific public policies for specific areas depending on the geographical and sociodemographic factors affecting the people who live there.
A data collection and analysis tool that works across agencies creates what Dr. Rodriguez calls a “public value intelligence”—the social and governmental equivalent of business intelligence—that is instrumental in designing and evaluating public policies using data the state generates.
To learn more about how cities are facilitating better planning and decision making, visit Microsoft CityNext’s Educated Cities website.
To read more about Azure solutions like the one the Ministry of Education is using, visit our Cortana Intelligence Suite website.
You can learn more about how Dr. Rodriguez is using advanced analytics and BI to enhance urban planning and educational transformation by viewing the TEDx talk in Spanish he gave at TEDx UniSabana in May 2016 or review the Youtube video in English.
