Published August 5, 2024
Newly minted urban planning doctorate Fuzhen Yin is carving out new territory in the field of urban informatics through research that links social media, machine learning and spatial analysis to generate insights on disaster response in cities.
The UB Presidential Fellow, a native of China with master’s degrees in landscape design and data science, came to UB with an interest in shrinking cities and urban informatics. That line of inquiry took a natural turn toward emergency response after she arrived in Buffalo in Fall 2020 – during the depths of the Covid-19 pandemic – and then experienced the Buffalo Blizzard of 2022.
Yin argues that with the proliferation of information technologies, human dynamics in cities must be explored across hybrid spaces, as activities in virtual spaces shape those in physical spaces, and vice versa. “It is important to bring data and measurability to these emerging dynamics to better inform urban and disaster response policy decisions,” says Yin, who first became interested in crisis planning after a flood crippled her hometown of Shinazhuang in 2016.
“While numerous studies in the field have examined the role of social media in community resilience, I wanted to look at how social media and, more broadly, cyberspace, can contribute to disaster responses.”
For her dissertation, Yin applies a space-place – or “splatial” framework – to the Buffalo Blizzard by mapping the direct impacts of social media interactions in generating mutual aid in physical spaces in the form of material, social and emotional support for thousands of Buffalonians.
The historic storm, which hit the city on Christmas Day 2022, dropped four feet of snow, paralyzed the region for nearly a week and caused 47 deaths. As the storm set in on Christmas Eve, citizens activated two significant Facebook Groups – Buffalo S.T.O.R.M. and Buffalo Blizzard Response – connecting more than 80,000 citizens and generating over 10,000 posts to share information, exchange resources and coordinate collective behaviors such as donations, wellness checks and memorials.
Applying machine learning, Yin coded a selection of the 10,000 posts into four categories – provide support, request support, express emotional support, and other – and then “trained” an algorithm to conduct a full content analysis of each post, comment or reply. Those data were then mapped to assess the structure of the network, actors’ roles within the network, how one’s position in the network affected aid distribution, and the degree to which requests translated into material support and mutual aid.
Nearly 40 percent of all messages – posts, replies and comments – came in the form of emotional support – prayers, expressions of gratitude and sympathy, and messages of encouragement, often expressed with emojis – while 30 percent of messages were offers of support and 7 percent were requests for help.
Yin, who says she received countless messages of support herself from friends and colleagues at UB as she experienced her first-ever blizzard – says her research reinforces that social media is effective in creating community resilience – rather than simply “making visible” existing resilience.
Consider that the Buffalo Blizzard Facebook Groups were particularly effective in introducing citizens on the periphery of established networks to sources of mutual aid as well as sustained communities of social support.
Specifically, more than two out of every three users had “weak ties” to the group – with fewer connections and interactions than more established users. Yin describes this as a “sparse network.” At the same time, Yin’s spatial analysis of interactions shows a flurry of activity around these new users, suggesting a high level of efficiency in linking new users to those with deep connections to resources.
“So many of these stories are heartbreaking, but at the same time they warmed my heart. This study brings measurability to the kindness and solidarity of this ‘City of Good Neighbors,’ said Yin, who was first introduced to Buffalo’s extensive grassroots community as a research assistant with the School’s Food Systems Planning and Healthy Communities Lab, which is closely connected to community actors in urban farming and food resilience.
Yin also said the sheer number of heartfelt expressions of emotional support indicate a critical indicator of valuable social capital – the intention and motivation to support fellow citizens. This may suggest the potential for future disaster response capacity that could be empowered by awareness-building among citizens.
During her tenure with UB, Yin was a prolific scholar, publishing six co-authored articles that apply urban analytics to food systems planning, smart transportation systems, and shrinking city issues like property abandonment and demolition planning. Yin is the first author in an article published in the International Journal of Geographical Information Science (April 2024), in which she extends another line of her dissertation research on the relationship of hybrid spaces and information propagation to Covid-19 vaccine uptake levels across New York State.
Faculty advisors to Yin’s dissertation were Lucie Laurian, professor and chair of UB’s Department of Urban and Regional Planning and an expert in urban governance and disaster preparedness; Emmanuel Frimpong Boamah, associate professor of urban planning who employs social network analysis and modeling in his research on planning processes and institutional structures; and UB geology professor Andrew Crooks, who specializes in Geographic Information Systems.
Laurian says Yin's work impacts several disciplines and fields of study: "Her work makes major contributions to the fields of urban planning, disaster management and response and resilience, and epidemiology, as well as significant methodological contributions in the areas of AI tool applications to social media networks and text (and emoji) analysis. She has already published several papers, and her dissertation work will translate into three to five additional papers that will establish her leadership in the field."
Yin added that the camaraderie and support of her faculty and doctoral cohort were fundamental to her success in the program, providing mentorship, direct support in the design of her research, and assistance with publication-writing and networking for employment.
The urban planning scholar heads now to the faculty of the University of Colorado, Colorado Springs, where she will extend her research by examining the long-term term impacts of social media interactions on community resilience, assessing network dynamics across different disaster types, and developing social media-integrated tools and software to support more efficient disaster responses.