Patricia J. Culligan

Robert, A. W. and Christine S. Carlton Professor of Civil Engineering

School of Engineering and Applied Science, Columbia University




A leader in the field of water resources and urban sustainability, Patricia Culligan explores novel, interdisciplinary solutions to the challenges of urbanization with a particular emphasis on the City of New York. Her research investigates the opportunities for green infrastructure, social networks, and advanced measurement and sensing technologies to improve urban water, energy, and environmental management. She is co-Director of a $12 million research network sponsored by the U.S. National Science Foundation (NSF) to develop new models for urban infrastructure to make cities cleaner, healthier, and more enjoyable places to live.  She is a faculty member of the Earth Institute at Columbia University and was the founding associate director of Columbia University’s Data Science Institute. She is the author or co-author of more than 160 technical articles. Dr. Culligan received her M.Phil. and Ph.D. from the University of Cambridge, England

Example Publications

Ramaswami, A. A. G. Russell, P. J. Culligan, K. R. Sharma, E. Kumar, Meta-principles for developing smart, sustainable, and healthy cities, Science, Vol 352, Issue 6288, pp 940-943, 2016.

Abrol, S., A. Mehmani, M. Kerman, C.J. Meinrenken, P. J. Culligan, Data-Enabled Building Energy Savings (D-E-BES), Proceedings of the IEEE, Vol. 106, issue 4, April 2018. DOI: 10.1109/JPROC.2018.2791405

Jain, R., R. Gulbinas, J. Taylor, and P. Culligan, Can social influence drive energy savings? Detecting the impact of social influence on the energy consumption behavior of networked users exposed to normative eco-feedback, Energy and Buildings, Vol 66, pp 119-127, 2013. Smart cities are defined as “smart, sustainable, healthy”. What about other criteria like livability, or relationships between members within the communities? Could you give us examples of projects where the planners took these questions into consideration and had to compromise?

Patricia J. Culligan: I would argue that the smartest cities of the future are going to be those that have the highest standards of livability, as defined by their inhabitants. These are the cities that will prosper, because they will be attractive places to both live and work. Important aspects of livability can include affordability, accessibility to services, vibrant public spaces and cohesive neighborhoods, among many other features. So by my definition, a “smart city” should be a city where data are used to generate information and knowledge that can be translated into actions that improve livability for as many citizens as possible. One example of a smart-cities project that is focused on best outcomes for citizens and involves a colleague of mine, Dr. Jacqueline Klopp, concerns urban mobility in the digital age. The Digital Matatus project[1] in Nairobi Kenya designed a new transit map for the city based on open source data collected from the informal network of privately owned mini-vans – matatus – which service millions of Nairobi citizens every day. A mobile phone based version of this map now makes it possible for these citizens to better navigate their city using the extensive matatus network. The alternative of introducing a formally planned, government run bus service would not only have compromised the livelihood of the matatus owners, it would also have reduced transit coverage, making it harder for the citizens of Nairobi to access services and jobs. Can you describe the type of resistance you have observed in projects that you were involved in? What are in your opinion, the root causes of this resistance and of attachment to status quo? What role do factors like culture, education, generation, or lack of trust in planning authorities play in creating such tensions?

Patricia J. Culligan: New York City’s green infrastructure plan, which was considered to provide a smarter and more sustainable alternative to the City’s stormwater management problems than a traditional grey infrastructure approach, is a good example of a project that met unexpected resistance. There was an assumption that placing engineered green infrastructure, such as rain-gardens, in the public right of way, would be welcomed by communities because of the environmental, and supposedly aesthetic benefits provided by green infrastructure. Instead, citizens expressed concern that the rain-gardens obstructed the public right of way. In addition, they did not like the native vegetation planted in the rain gardens, and could not understand why the City was investing in this infrastructure when they perceived a need for other infrastructure investments in their neighborhoods. As a result of resistance to this project, rain-gardens have been vandalized, which has compromised their performance. I believe that the root cause of this resistance was a decision to impose an, albeit smarter – but very visible – solution to an urban problem on a city’s citizens. Attachment to the status quo, as well as lack of information and education about the services provided by green infrastructure were also factors. But ultimately, smart cities need to use the extensive information and communication technologies currently available to engage citizens as broadly as possible in decision making. Top-down approaches, even if they are well-intended, are likely to meet resistance in today’s connected world. Smart cities require access to enormous amounts of personal data. Are concerns about privacy in relation to such uses of personal data exaggerated or is it a real issue? What is concretely at risk? Who are the people the most concerned? What can you do on your end to respond to these concerns?

Patricia J. Culligan: I see an interesting generational divide with respect to concerns about personal data and privacy. I feel that those of us who have not grown up in a digital age have greater concern than the youth of today, who enjoy sharing multiple forms of personal data, including their daily activities, likes, dislikes and 24/7 location. To some extent, I feel that many of the youth of today have accepted that the advantages of today’s data-driven society come with a reduction in personal privacy.

The question as to whether there is a real risk associated with the wealth of personal information gathered about individuals today, rests on how that information is used. Information that is anonymized and used solely to enhance the experiences of citizens would not appear to carry much risk. However, information that is not anonymized and is used to make decisions that impact individuals or communities – especially without their knowledge or input – has great risk of fostering discrimination. My own biggest concerns rest around the anonymization and security of personal data, and whether the curator of important data sets is responsible, transparent and ethical. I think a large number of cities, including New York City, have shown great leadership in open publication of anonymized data sets that have been collected or generated using public funds, and these data sets have become valuable resources to many, including the city’s citizens themselves. When cities contract data collection and analysis to the private sector, however, I become concerned about what mechanisms are in place for anonymization, data security, sharing and transparency. We have seen too many recent cases of the private sector abusing the personal data or information they have collected.  We hear a lot about algorithms interpreting mass of data and identifying patterns. Are you satisfied with the end results of these analytics? Do you feel like there are things that algorithms are unable to take into account, such as interpretations of causal relationships? Should algorithm-based research be accompanied by other, more intersubjective forms of research? Is this something that you are doing?

Patricia J. Culligan: Algorithms can be very effective for detecting patterns in large data sets, although end results are dependent on how the data set was gathered and down-sampled, in cases where down-sampling is necessary. I think the bigger question relates to the use of these patterns for forecasting and decision making. Here is where I think there can be an over-confidence in, or misuse of such patterns, because in many cases factors causing the patterns are unknown. For example, over this past year, my colleagues and I have gathered a very large data set on electrical consumption in over 400 apartments located in New York City, which we have collected using unobtrusive smart-metering techniques. We have used patterns in the data set to quantify the average minimum electrical consumption in each apartment. We call this the “Vampire Load”, and we are interested in it because a reduction of this load will lead to big, overall savings in urban energy consumption. Interestingly, we have found a very large variation in the average minimum energy used, per square meter, across the 400 apartments. However, what our data set can’t tell us, is why this variation exists. To understand this, we need supplemental information on how many people occupy each apartment, the appliances they own and use and how they manage and operate them. To collect this information, we need to use traditional survey techniques. We also can’t use our data set to forecast minimum energy consumption per apartment going into the future, as the data set only applies to current conditions. Apartment dwellers might, for example, increase their future vampire load via the purchase of additional appliances, or they might decrease their load by replacing an existing appliance with a newer, more efficient model. One of the biggest mistake people make in using patterns in large data sets for forecasting purposes, is lack of recognition that the conditions under which many data sets are collected can change with time and setting.  Can you give us examples where social dynamics have given a significant momentum to the smart cities projects? 

Patricia J. Culligan: The work my colleagues and I have been doing on motivating energy conservation behavior for city apartment dwellers, has confirmed that providing people feedback on their energy consumption motivates them to reduce consumption. However, we have also found that providing feedback to a social network on every member’s individual consumption, motivates even greater savings. Some of this has to do with sharing of tips on how to conserve energy across the group, but a lot of it has to do with social dynamics. In particular, if social group leaders or influencers adopt energy saving behaviors, the overall savings of the group increases. As human connectivity increases, the role of social group leaders and influencers in supporting, or otherwise, momentum towards smart cities approaches that improves livability for citizens, is going to be very important.

Propos recueillis par Pierre Bismuth

« How can social dynamics help make cities smarter », an interview with Dr Patricia Culligan, a leader in the field of water resources and urban sustainability, Columbia University



Pierre Bismuth