Big Data in the Big City

The era of big data in public policy has begun, and the promises and limitations of this social science tool will soon be tested in cities. Policy outcomes can now be measured with greater accuracy and finer location than ever, improving the delivery of urban services and accounting for their value to society.

Alas, big data cannot solve urban social problems on its own. In a new working paper, Edward Glaeser, Scott Duke Kominers, Michael Luca, and Nikhil Naik review when and how big data can have an important policy impact and how information from the Internet is being used to improve city management.

From online social networks to the computerization of public records, big data can help answer urban questions. Big data is also improving city management, as municipal operations, local regulations, the allocation of scarce resources, and financial forecasts can all benefit from data-driven evidence. For example, Seattle has partnered with Accenture and Microsoft to track energy usage in public buildings downtown and has developed algorithms to cut consumption and cost.

Connecting Physical Cities to Social Outcomes

Planners and architects have long promoted the idea that physical space helps shape social outcomes. Data from GPS imagery like Google Street View can now be used to measure urban spaces and their impact on social outcomes. For instances, Nikhil Naik, study co-author and part of the MIT Media Lab, created StreetScore as an algorithm to estimate perceived safety of different streets. Measuring street images could also enable research into patterns of poverty, segregation, and development.

Estimating the Value of Urban Amenities

Policymakers often estimate the value of urban investments by weighing the social costs and benefits of a given policy. These estimates can be enhanced by big data. Two methods of estimating social costs and benefits involve revealed preferences and contingent valuation.

Revealed preferences refers to inferring consumer value for a non-market good through the prices of other goods. For instance, hedonic housing price models attempt to capture the ‘price’ of fresh air by looking at housing prices in areas that have or lack fresh air. These models could be augmented with Zillow data on home prices and amenities to create more accurate price data, and to inform policymakers of the value of public amenities. Cities could also improve the appraisal process for property tax purposes by using predictive models that incorporate this type of big data.

The second method, contingent valuation, is often used by environmentalists. This technique surveys respondents about their willingness to pay for an environmental amenity, like a park or endangered species. To boost the value of survey data for urban use, researchers need to make choices comparable and compatible with personal experiences. For example, to evaluate a new park in Chicago’s Northerly Island, one could ask residents if they prefer a new park or that the revenue from developing that land go toward renovating existing parks.

Mining social networks, like Twitter and Facebook, for public sentiment related to urban amenities can also help policymakers approximate valuation and revealed preferences. Though these findings would not be as robust as those elicited from a survey written to extract valuation of, say, a public park, sentiment analysis from social networks could still be a useful tool in the valuation toolbox.

Improving the Quality of City Services

Perhaps the best use of big data in city management is through the provision of better services. The New York City Police Department targets resources by mapping models of criminal activity to increase efficacy and accountability. In Boston, the mobile app Street Bump enables residents to report needed street repairs and allows the public to monitor progress. Glaeser proposes aggregating Yelp ratings through word searches for “dirty” and “sick” to help allocate health inspectors that would cut cost and promote public health.

Big Opportunities for Big Data

Measurement is often a barrier to better policy that can be overcome with the help of big data. Street imaging, compatible constituent surveys, and direct enhancement services can all be effective in helping to convince policymakers that fundamental changes are needed in municipal governance. In this era, using data-driven evidence to determine city investments can improve urban life.

Article Source: Glaeser, Edward L., Scott Duke Kominers, Michael Luca, and Nikhil Naik. “Big Data and Big Cities: The Promises and Limitations of Improved Measures of Urban Life.” Harvard Business School NOM Unit Working Paper 16-065, 2015.

Featured Photo: cc/(NicoElNino, photo ID: 76622851, from iStock by Getty Images)