Researchers Track Twitter to Learn What People Value in New York City Parks
For Immediate Release
North Carolina State University researchers found they could use Twitter to understand changes in what New York City park users valued most about four iconic city parks before and after COVID-19 lockdowns went into effect. The researchers also found Twitter useful for tracking complaints about individual parks.
The findings indicate social media could be an important tool for park managers, park planners or others to respond in real time to changes in users’ needs or to plan for future parks, potentially faster than using traditional survey-based methods.
“We found that you can pull out detail about individual parks, as well as track what people value in parks, complaints they had about specific events, and even broader societal issues that people are talking about,” said study co-author Aaron Hipp, associate professor of community health and sustainability at NC State. “While we have some additional work to automate this and get closer to real-time monitoring, we think our findings indicate that parks can monitor this information and feel confident that some of the social media traffic can be a pretty reliable reflection of public sentiments.”
For the study, researchers tracked tweets about Central Park in Manhattan, Prospect Park in Brooklyn, Flushing Meadows in Queens and the Bronx Park from March 2019 to February 2020, and then from March 2020 to February 2021. They used natural language processing and topic modeling techniques to analyze a total of 71,792 tweets, which excluded retweets, tweets from influencers, bots, and tweets containing fewer than three words. Overall, they found tweets about Central Park grew 18% and tweets about Prospect Park grew by 87%. Meanwhile, tweets about Flushing Meadows declined 71% and tweets about Bronx Park fell 32%.
Researchers said the findings reflect the effects of a COVID-19 stay-at-home order. Since outdoor recreation was limited to solitary exercise, parks and trails became popular destinations after March 2020. However, this trend did not seem to hold true across all parks due to their distinct offerings.
“Flushing Meadows is famous for sports events, but during the pandemic, those were mostly canceled,” said the study’s lead author Jing-Huei Huang, postdoctoral research scholar at NC State. “In Bronx Park, people tended to tweet about the zoo and Botanical Garden, but those were closed, so that was reflected in the decline in tweets.”
Tweets about physical activity such as walking, jogging and biking grew in all four parks during the pandemic. The researchers also saw concerns about social distancing and related issues in all four.
“In all four parks, we saw participating in physical activity outdoors was particularly important when people had to keep distant from each other, and they were not able to participate in social activities,” Huang said. “We were also able to capture negative feelings when people were upset seeing crowds in parks, or when people weren’t wearing masks,” Huang added. “It’s sending useful signals to the management team.”
They also captured tweets about specific incidents in each park that reflected larger social concerns. For example, they saw a peak in tweets about racial discrimination in May 2020 after the Central Park birdwatching incident, when Amy Cooper, a white woman, called the police on a Black man, Christian Cooper, who was birdwatching. They also saw complaints about Flushing Meadows becoming a “giant parking lot” amid advocacy for a new greenway to increase access and ensure pedestrian safety.
In a follow-up study, they’re planning to compare data gathered using traditional surveys to what they found on social media.
“A long-range goal is to be able to get this feedback in real time, so park managers could launch programming in response,” Hipp said. “For example, amid a surge in demand, you could open a street nearby for waking or biking. There is also big potential here for the evaluation of programs and events especially when decisions are made on the fly, such as during COVID. This social media data can provide a retrospective opportunity to evaluate what people were saying about the park, a specific program or event.”
The paper, “Exploring values through Twitter data associated with urban parks pre- and post-COVID-19,” was published online in Landscape and Urban Planning. Co-authors included Myron F. Floyd and Laura G. Tateosian. The study is part of the project “Greenspace Characteristics and their Associations with Population Health” funded by USDA Forest Service Grant/Agreement Number: 16-JV-11330144-065
Note to authors: The abstract follows.
“Exploring values through Twitter data associated with urban parks pre- and post-COVID-19”
Authors: Jing-Huei Huang, Myron F. Floyd, Laura G. Tateosian and J. Aaron Hipp
Published online in Landscape and Urban Planning.
Abstract: Since school and business closures due to the evolving COVID-19 outbreak, urban parks have been a popular destination, offering spaces for daily fitness activities and an escape from the home environment. There is a need for evidence for parks and recreation departments and agencies to base decisions when adapting policies in response to the rapid change in demand and preferences during the pandemic. The application of social media data analytic techniques permits a qualitative and quantitative big-data approach to gain unobtrusive and prompt insights on how parks are valued. This study investigates how public values associated with NYC parks has shifted between pre- COVID (i.e., from March 2019 to February 2020) and postCOVID (i.e., from March 2020 to February 2021) through a social media microblogging platform –Twitter. A topic modeling technique for short text identified common traits of the changes in Twitter topics regarding impressions and values associated with the parks over two years. While the NYC lockdown resulted in much fewer social activities in parks, some parks continued to be valued for physical activity and nature contact during the pandemic. Concerns about people not keeping physical distance arose in parks where frequent human interactions and crowding seemed to cause a higher probability of the coronavirus transmission. This study demonstrates social media data could be used to capture park values and be specific per park. Results could inform park management during disruptions when use is altered and the needs of the public may be changing.