Exploring Urban Science

Ever increasing sources, quality, and types of data translate into a new opportunity to leverage data science and analytics to help city planners, infrastructure engineers, and policy-makers both understand and engage with cities in entirely new manners. But how do we merge the imperfect, real-world datasets urban planners must work with and the classical computer science training, that seeks to optimize outcomes through careful control of polished datasets? Yuan Lai, Lecturer of Urban Science and Planning, hopes to collaborate with MIT’s 11-6 undergraduates to answer questions like these and, in the process, better define the field of urban science.

Prior to coming to MIT, Lai was a research affiliate at NYU Marron Institute of Urban Management and the NYU Center for Urban Science and Progress (CUSP), where he utilized applied analytics and machine learning to combine high volumes and disparate types of data into actionable research that could be used to improve the health of NYC residents and facilitate urban planners’ understanding of changes in the urban fabric of a city. Lai also practiced architecture and urban design at Safdie Architects, where he worked on large scale mixed-use development projects worldwide. He holds a PhD in Civil Engineering with a concentration in urban systems and informatics from NYU, a MS in Applied Urban Science and Informatics from NYU CUSP, a Master of Urban Planning with concentrations in GIS and urban design from SUNY Buffalo, and a Bachelor of Landscape Architecture from Beijing Forestry University.

Q1. Often, when one conjures a mental image of ‘smart cities’ we think of cities that are using sensors to provide more – and more complete – data. Is there an alternative model of expanding urban science?

Yuan: Cities nowadays maintain vast information relevant to planning, management, and infrastructure investment. Sensor-generated data represents an important part of the urban data landscape, along with digital inventory, administrative records, economic transactions, social media, and others. Increasing public access to such data can provide new sources for generating useful applications and new insights. However, most urban data represent as "digital exhaust", where the potential uses are often far beyond the original rationale for collecting the data. For example, my recent publication in Computers, Environment and Urban Systems presents a data mining and machine learning process for extracting, analyzing, and integrating building permits from more than 2.5 million building alteration projects from seven major U.S. cities. This approach may assist cities to better monitor building activity, analyze spatiotemporal patterns of development, and more fully understand the economic, social, and environmental implications of changes to the urban built environment. It contributes to a methodological foundation for near-real-time construction activity analytics as a complementary approach to classify construction activity while recognizing each city's administrative and technical context. This exploratory study, potentially as a part of urban science, contributes to the growing applied data science related to urban planning and city management.

Q2. Urban Science is a relatively nascent field of study, even the definition remains elusive to many urban planners. How do you define it?

Yuan: The breadth and complexity of cities require diverse domain expertise and cross-cutting collaborations. I do believe that faculty at DUSP, in collaboration with colleagues from other departments at MIT, collectively define urban science. In general, I tend to describe urban science with its purpose, process, and product rather than a fixed definition. In the recent two decades, information and communication technologies and the Internet of Things has integrated with the urban environment at an unprecedented speed. While data as intermediate layers connecting urban life with promises, it also brings new conflicts and uncertainties. As data is becoming more accessible it also enhances the data’s power of shaping decisions and nudging behaviors. Therefore, it is increasingly important to utilize data with domain knowledge, scientific methods, validation, and ethics. Since urbanization and computation are both major forces shaping our future, it is critical for the younger generation to be capable of thinking and working on both aspects.

Students in urban science will deeply engage in developing the understandings and skill-sets for handling socio-technical complexities in cities. Urban science is not just about analyzing data but using the city as a living laboratory to explore solutions with a meaningful purpose, tangible use, and hopefully positive impact. These goals cannot rely on an ad-hoc effort in a silo, but a collaborative environment with interdisciplinary research and local community’s engagement. The talented students, diverse faculties, and wide array of resources at MIT, as well as the general innovative atmosphere in Boston/Cambridge area make it the right environment for exploring and codifying the field of urban science.

Q3. Could you speak more to local, tangible products that can be implemented by the MIT community?

Yuan: I am very interested in combining computing and urban research methods to create tangible applications for better quality-of-life at local scale. One of my previous projects published in Health & Place, as an example, investigated street tree species in New York City and their pollen’s impact on local air quality and respiratory health. This study illustrates how cross-domain data integration at high spatial resolution can address complex environmental justice issues by quantifying local environmental, land use, and socio-economic characteristics. The availability of data resources from numerous activities in cities provides new opportunities for improved urban decision-making and more efficient and equitable services for citizens. In addition to the hypothesis testing, these data integration methods can be leveraged to inform residents through information visualizations or a location-based mobile application.

We hope 11-6 can attract students who are interested in computing and prototyping, but also care about urban issues related to sustainability, transportation, equity, health, and economic development. One possible future project is to explore innovations for more active, social, and healthy living through a combined computing and urban studies methods. Students will collect data, design analysis, and create prototypes that may have a real-world implementation in MIT community, possibly in an extended Cambridge area. Students can apply the programming and analytical skills they have learned from computer science courses to explore possibilities for better design, planning, operation, and policy in urban context. Through this process, they will also understand the potential social-ecological-economic consequences of technology deployment and data-driven decision-making in reality.