Aniruddha Dasgupta, World Resources Institute
Jeanne Holm, City of Los Angeles
Sallie Keller, University of Virginia
Katherine Bennett Ensor, Rice University
Katherine Bennett Ensor, Rice University, emphasized that collaborations between researchers and practitioners or between universities and cities are beneficial; the co-creation of data, knowledge, and models produces a shared resource that advances all interested parties. She championed the cohort model for larger communities that span many municipalities or for a network of smaller communities. She added that these partnerships are particularly advantageous when disasters and unusual incidents occur, because a larger team is ready to help face those together. Collaborations can bring in better expertise more quickly for the immediate and long-term recovery from such events.
Jeanne Holm, City of Los Angeles, said that cities have limited bandwidth for partnering with universities. A regional association of governments is a powerful tool that worked well in Los Angeles. Any cohort, federation, or collaborative should encompass more than just universities
(e.g., city colleges, Girls Who Code1), and she encouraged people to be creative and inclusive in how they approach collaboration for urban sustainability—for example, through storytelling initiatives, qualitative data gathering, and experiential discussions. It is also valuable to consider students’ experiences in cohorts, which can provide good exposure to civil service careers and allow students to apply their education in a meaningful way. She added that it is vital to have a feedback loop connecting theory, policies, actions, and outcomes; this gives information back to the research community to avoid the development of uninformed theories. Holm suggested that the academic perspective move from finding “the shining star of truth” to developing a cycle of learning, iterating, and contributing. She said that although truth can sometimes be found, differences between cities affect how policies are implemented. She stressed that it is vital to think globally and inclusively, following the examples of organizations such as 100 Resilient Cities and the World Resources Institute (WRI). Auroop Ganguly, Northeastern University, agreed with Holm’s perspective. He added that there may be unintended consequences of overconfident academic researchers and emphasized the value of the iterative process. Sallie Keller, University of Virginia, underscored that industry and the commercial sector should be a part of any partnership discussion. Community stakeholders should be engaged deeply in data-driven community research and decision making, instead of only being recipients of the academic research and student projects.
Aniruddha Dasgupta, WRI, reiterated the theme of the university–city partnership that was prevalent throughout the workshop. He highlighted the localization of knowledge and expressed his hope that progress that has been made in certain cities can be replicated in other cities. He wondered what incentive structures need to change in academia to recognize practice-oriented work and how cities could develop incentives to learn to work better with academia. Anu Ramaswami, University of Minnesota, proposed that new models be considered to increase the longevity of partnerships, such as the large networks supported by the National Science Foundation. She encouraged participants to be entrepreneurial and to continue to engage universities and cities.
Gyami Shrestha, U.S. Carbon Cycle Science Program (a federal inter-agency partnership), described her organization’s decadal state of the carbon cycle science report, which assessed carbon cycle science across land, air, water, and society in North America. She said that urban sustainability will be attractive for many years and encouraged everyone to consider how to enhance the feedback loop and interactions among
universities, cities, states, federal funders, and international funders. She also suggested that existing funding and activities be leveraged to achieve this common goal (e.g., the National Aeronautics and Space Administration has funded much research on megacities).
Data Access, Ethics, and Privacy Issues
Keller said that improved strategies are needed to distribute information about community resources. It is important that problem solving remains the focus during the data discovery process. She emphasized the value of bridging agency stovepipes through data and exposing people to the wealth of survey and administrative data resources available at local, state, and federal levels as well as Internet-based data. There are not many off-the-shelf products available for data analytics for urban sustainability, but valuable analytics are emerging through the use of sophisticated statistical methods. A body of peer-reviewed research around community problems is also emerging.
Ensor emphasized the value of data archiving and curation; data used to make decisions should be permanently present. Holm advised against reinventing data platforms. Instead, data.gov is an example of an established platform (including 8,000 data sets from cities) that researchers find useful. However, she noted that even this platform is not immune to issues of permanence (e.g., government shutdowns would make this platform inaccessible). She said that maintaining this platform should be considered an essential service. She echoed Keller’s assertion that a thorough data discovery is necessary to understand what data are available to address a research question. She emphasized the value of citizen science and crowdsourcing—people with mobile phones and wearable devices are constantly generating data, and she hopes that this work is lifted to a level of rigor in which it can be used in academic research.
Data privacy and ethics are important components of data access. Keller expects that there will be confusion in communities about privacy and whether data can be used for certain circumstances, so more education and conversation about both privacy and ethics are needed. The institutional review board (IRB) structure is well known in parts of academia and industry, but the public and local governments are not necessarily aware of these processes. The IRB framework could help the public to better understand informed consent and privacy, as well as ethical considerations around who is benefiting from a particular analysis. Conversations about ethics and privacy as well as about the value of transparency and reproducibility need to play a prominent role in both partnerships and policy making, Keller asserted.
Dasgupta said that as more data are generated and stored by private entities, ethical issues will continue to arise, especially as those data assets are used publicly. He echoed Keller’s suggestion for the generation of rules inspired by IRB standards. Audience member David Rabinowitz questioned what happens with individuals’ data. Rather than making data available only for general policy generation, they could be made available to the individuals and perhaps used as motivation to change behavior (e.g., data from utility companies about how much electricity an individual uses as compared to her neighbor). Keller noted the value of this suggestion and added that a different definition of “smart city” is needed: a city is smart if the individuals in a city can access the information they need, when they need it, to improve their lives. As an example, Holm mentioned that through a partnership with Waze, the city of Los Angeles shares information about high traffic-accident injury networks with the app so that it can direct people around potential accident spots. She described this as a highly synergistic way to share information.
Will Angel, NaskMe, asked Holm to elaborate on the successes of such partnerships because, at the city level, larger companies are sometimes reticent to share their data. Holm explained that companies like Waze do not particularly want to share their data, and Los Angeles does not want to access individual data, so there has to be a negotiation with each company (each of which will have a slightly different business model) to try to find the place in which both the city and the company are willing to share information to prevent public harm. She highlighted the National Association of City Transportation Officials as an appropriate place to have this transportation-data-focused conversation about balancing individual companies’ interests in protecting data with cities’ desire to be better informed by that data. Luís Bettencourt, University of Chicago, said that verification, transparency, and curation become even more important with an increase in available data. It is important to understand which signals can be verified and have the potential to align entities around solutions. The context around data matters, as does integrating data to achieve public good.
Ganguly hoped that people at this workshop would form partnerships, as participants have expertise in data, context, and solutions. He also suggested crowdsourcing efforts based on available data and models that could be built. David Maier, Portland State University, endorsed the idea of coming up with benchmarks with ground truth, both to test the models and to test the sufficiency of data.
Better Decision Making
Ensor asserted that better decisions happen in partnership. Researchers’ methods need to align with the problem and the partnership without compromising robustness and validity. She emphasized that decisions in the sustainability framework are multifaceted, which is why a systems approach is essential. Although it is not always possible to arrive at a perfect answer, it is crucial to have a robust answer with a transparent methodology, Ensor continued.
Holm said that from the perspective of a city employee, transparent decision making is enabled by confidence that it is safe to make data available and that cities (or their employees) will not be judged if those data are imperfect. She explained that theories and academic research in cities need to be implemented in a way that civil and public servants do not have to sacrifice their jobs or reputations. She also noted that academic researchers need to articulate how their theories could be actionable and make a difference. Ganguly asked about how academic researchers can best translate information, as some audiences will desire more information than others. Ensor explained that it is important to bring the correct level of science to the conversation, and Holm noted that city employees can have very different backgrounds, so that not all of them are savvy with data or scientific research. She suggested that a chief data/technology officer who could translate back to the political spectrum would be an asset in any city department. Holm said that this is also the best person to reach out to for any conversations about research and action-oriented decision making. Dasgupta commented that being inclusive of different points of view in the data generation and discovery process will provide a multidimensional view of reality and truth, which could lead to multiple valid solutions.
Capability of Cities to Use Data
Dasgupta pointed out that the majority of people in the world do not live in big cities that are capable of using complex data. He reiterated that a gap exists between operational-minded people who run cities with budget optimization in mind and those who use data to make decisions related to issues of sustainability. Holm noted that even when city officials are not tech savvy, most are interested in making better decisions for their residents, businesses, and visitors. Thus, it is important to identify existing political barriers to data-driven decision making and to think about how private-sector funding could offset budgetary constraints. Even though there will be turnover in the positions of elected officials, she explained that city departments will endure (i.e., garbage will always
need to be collected, and buildings will always need to be painted). Staff in those departments might be more willing to try new things, and often there is a political desire to partner with people who can provide data to make better decisions.
Ensor said that new statistical methodologies allow people to work across private databases without having to share the data; this is helpful when the goal is to get to a decision supported by information, not necessarily to hold the data. Keller added that there is a growing body of technology and solutions for issues related to data governance. Maier asked which of the models discussed throughout the workshop could assist cities with operations or short-term decision making. Keller mentioned that Bryan Lewis’s work at the University of Virginia relies on nowcasting for real-time decision making. Holm added that from a city government perspective, nowcasting is crucial for emergency management and, more generally, would be useful to understand how to better serve residents. Ensor said that a systems approach to modeling would be more useful for longer-term planning but that nowcasting has value for short-term decisions. She added that a future workshop more specifically about modeling could be useful in addressing this question more deeply.
Ulrike Passe, Iowa State University, explained that as an architect she focuses on the power of design, which is the link between data and decision making. She noted her surprise at how few designers attend meetings on urban sustainability. Keller said that new technologies are emerging to help move from a planner’s static view of the city to something more dynamic and to engage people (in all phases and parts of the research) who might be affected by these decisions. Holm emphasized the value of virtual reality because a blueprint is so different from a walk-through of a physical space. If a goal is to design a future city that will elevate its residents’ data literacy, then data has to be portrayed in a consumable way. Rabinowitz referred to the Smithsonian Museums’ strategy to tell stories with artifacts, and he stressed that the best way to communicate data is through storytelling. Holm agreed and said that it is also important to recognize the data divide and encourage people to understand how data could change their lives for good or bad through the sharing of stories—open source, open data, and open science are all part of achieving data literacy, and GitHub is an example of a platform that makes that achievable.
Keller hypothesized that a conversation at a similar workshop 10 years from now would be very different because these ideas will have been infused in the workforce and in local government. She suggested that communities begin to think now about future data innovations for governance and applications. Ensor highlighted efforts under way at the
K–12 and 2-year college levels, as well as in the National Academies2 and the American Statistical Association, to improve data literacy. She encouraged striving to overcome the digital divide so that the society of the future is data literate. Holm identified the Computer Science for All initiative,3 Khan Academy’s Hour of Code,4 and various bootcamps as traditional and nontraditional ways of data science learning that have emerged. She suggested embracing those methods for the students of the future, bringing rigor into conversations about data literacy, and encouraging inclusivity of populations.
Michelle Schwalbe, National Academies of Sciences, Engineering, and Medicine, asked how the National Academies could aid in the continuation of these discussions or whether there are other key stakeholders with whom to engage. Ensor suggested resuming the conversation with a series of roundtables that drive both scientist and practitioner inquiry. Keller said that because this workshop’s discussion was U.S.-centric, the National Academies could help organize a forum that would better cover the global landscape (i.e., the Global North and Global South)—there is much to be learned at the global level. She also proposed an activity that focuses on the types of city partnerships discussed throughout this workshop. Holm expected that many projects could emerge from this workshop—for example, a marketplace for matchmaking between cities and universities. She hopes that the group can connect and collaborate on projects going forward. Another important topic for further discussion is understanding better ways of accessing data, which might be an area in which the National Academies could facilitate or collaborate.
Keller said that more people and more opportunities are needed to publish the work emerging from partnerships in peer-reviewed literature (instead of only in popular media). Christine Ehlig-Economides, University of Houston, highlighted the need for funding that incentivizes city involvement in partnerships. Bettencourt noted that people from the humanities, social sciences, and policy arenas offer a different approach to
2 These efforts include the Roundtable on Data Science Postsecondary Education (see www.nas.edu/dsert) and the 2018 report Data Science for Undergraduates: Opportunities and Options (see www.nap.edu/25104).
data, which should alter perspectives of science, technology, engineering, and mathematics education reform. Dasgupta wondered about strategies to build the National Academies’ partnerships with other academies throughout the world. He closed the workshop by thanking participants, noting that society is on the cusp of big changes and expressing his hope that the National Academies would continue this conversation.