Credit: Ryan Brandenberg

“When we have all data online, it will be great for humanity. It is a prerequisite to solving many problems that humankind faces.”

Robert Cailliau, Co-Founder, World Wide Web

Cities are growing at an uncontrollable rate. 68% of the world’s population is expected to live in urban areas by 2030. Given this fast and unstoppable growth, it’s important to understand how cities could be planned better. This spring, I moderated a panel discussion on smart urban planning driven by data featuring two experts: Sveta Milusheva, an economist in the development impact evaluation (DIME) at the World Bank, and Sunil Wattal, associate professor of management information system at the Fox School of Business. In this panel discussion, organized by the Translational Research Center, we discussed how different types of data sources could be used for smarter urban planning.

These are the four key highlights from our discussion.

Crowdsourced Data for City Planning. Crowdsourced data is based on a simple philosophy that crowds of people can generate information faster than a single individual can. Tweets about national disasters in any country can tell the policymakers about different factors impacting people in real-time, something not possible with traditional research tools, such as surveys.

Sveta Milusheva’s project on road safety highlights this point. Milusheva, along with her collaborators, applied a machine learning model to 874,588 traffic-related tweets extracted from a popular Twitter handle in Nairobi, Kenya. Analyzing these tweets led them to create the first interactive and real-time map showing where and when the crashes occur. As Milusheva’s work shows, analyzing data openly shared by the public can help authorities improve infrastructure planning.

Sharing Economy and City Planning. Sharing economy, built on the idea of sharing assets and services between private individuals, is expected to grow to $335 billion by 2025. How does sharing economy impact our cities? Sunil Wattal elucidates both the positive and negatives. His research on Uber highlights that sharing rides reduce driving fatalities. Examining publicly available data on road fatalities before and after Uber X entered California, Wattal found that Uber X services reduced motor vehicle-related deaths by 3.6% to 5.6%, which translates to saving 500 lives per year. Wattal elaborated, “This creates a public welfare net of about $1.33 billion.”

In other research, he examined the effects of Airbnb’s “One Host, One Home” policy implementation in New York and San Francisco on crime rates. Analyzing publicly available crime data before and after this policy implementation, Wattal and his colleagues found that a decrease in Airbnb listings resulting from the policy led to a reduction in assaults, robberies and burglaries, but increased theft incidents. The increase in thefts was limited to high-income neighborhoods. They also found that the entry of Airbnb reduced crime rate in low-income neighborhoods.

One reason for this could be that sharing economy creates employment opportunities. Wattal’s research suggests that policymakers need to dedicate resources to enhance safety in high-income neighborhoods where Airbnb is prevalent. They should also increase connectivity in the low-income neighborhoods to encourage more travelers to book Airbnbs in these areas.

Combining Data Sources: Milusheva argues one could also benefit from administrative data sources. For example, when studying traffic incidents in Naorobi, Sveta and her collaborators digitized over 10,000 police records of traffic fatalities; these records uncovered that 70% of the accidents involved pedestrians and predominantly happened in areas where the informal bus system was the major transport mode. Relying only on the Twitter data would have led researchers to miss these important points. However, combining the publicly available and administrative data can provide richer insights and help policymakers better direct their resources to reduce traffic fatalities.

Checking Data: Checks should be put in place to ensure that researchers are using this data in the most scientific and rigorous manner, rather than engaging in a fishing expedition. In our discussion, we outlined four suggestions.

• Make sure to start with a specific hypothesis before you start digging into the data. A proper understanding of the underlying psychology and having a theory-based conceptual model is crucial.
• Make sure that your results are not driven by outliers. With high-frequency data, researchers could at times get crazy peaks that happen by accident. Therefore, it is important to clean your data carefully to make sure your conclusions are not driven because of the outliers.
• Be aware of the limitations of data. For example, people are more likely to tweet during certain times in a day, thus making this data not very useful for other times.
• Finally, one must ensure the privacy of people. This could be done by aggregating data to look at useful patterns.

The new world we live in poses different challenges for urban planning. However, we have many different innovative data sources today, which both researchers and policymakers can use to come up with effective solutions. What drives interesting research insights is not the bits of data we have, rather how creatively we use different data sources.

Monica Wadhwa is an associate professor in the Fox School’s Department of Marketing and Supply Chain Management. Her research focuses on understanding the motivational and affective determinants of consumer decisions making. She is also the research impact director for the Translational Research Center.