The largest parks, greater than 100 acres, had the highest mean happiness benefit, possibly because that larger parks provide greater opportunities for mental restoration and separation from the taxing environment of the city

Schwartz AJ, Dodds PS, O’Neil-Dunne JPM, Ricketts TH, Danforth CM (2022) Gauging the happiness benefit of US urban parks through Twitter. PLoS ONE 17(3): e0261056. Mar 30, 2022. https://doi.org/10.1371/journal.pone.0261056

Abstract: The relationship between nature contact and mental well-being has received increasing attention in recent years. While a body of evidence has accumulated demonstrating a positive relationship between time in nature and mental well-being, there have been few studies comparing this relationship in different locations over long periods of time. In this study, we analyze over 1.5 million tweets to estimate a happiness benefit, the difference in expressed happiness between in- and out-of-park tweets, for the 25 largest cities in the US by population. People write happier words during park visits when compared with non-park user tweets collected around the same time. While the words people write are happier in parks on average and in most cities, we find considerable variation across cities. Tweets are happier in parks at all times of the day, week, and year, not just during the weekend or summer vacation. Across all cities, we find that the happiness benefit is highest in parks larger than 100 acres. Overall, our study suggests the happiness benefit associated with park visitation is on par with US holidays such as Thanksgiving and New Year’s Day.

Tweets inside of all park size categories exhibited a positive happiness benefit. The largest parks, greater than 100 acres, had the highest mean happiness benefit. One possible explanation is that larger parks provide greater opportunities for mental restoration and separation from the taxing environment of the city. This finding is consistent with results from our earlier study in San Francisco, in which tweets in the larger and greener Regional Parks had the highest happiness benefit [18]. Parks between 0 and 10 acres are often neighborhood parks that people use in their day to day lives. Local parks provide many essential functions; however, our results suggest that the experiences people have in larger parks may be more beneficial from a mental health perspective. Another possibility is that people spend more time in larger parks; one study suggested that 120 minutes of nature contact a week resulted in improved health and well-being [35].

Temporal analysis

Across all cities, we grouped park tweets and their control tweets according to the in-park tweet’s timestamp to test H4. First, we compared the happiness benefit by season. The mean happiness benefit was highest in the summer (0.12), followed by fall (0.10), spring (0.08), and winter (0.06) as shown in Fig 4B. Then we grouped park tweets and their respective control tweets according to the day of the week in which it was posted. Saturday exhibited the highest mean happiness benefit (.15) followed by Sunday (0.13). Monday through Friday were all between 0.06 and 0.09 (Fig 4). We also estimated the happiness benefit by hour of the day. The tweets posted during the 8:00 and 9:00 AM hours had a mean happiness benefit around 0.07 while the rest of the day did not show a clear pattern, ranging from 0.08 to 0.14 (S4 Fig).

We observe that the mean happiness benefit was higher in summer than other seasons; however, the happiness benefit was positive in all four seasons. Possible interpretations of seasonal differences may include that warmer or sunnier weather in the summer leads to an increased benefit from park visitation. People may engage in longer visits to parks during summer months, engage in physical activity, or connect with friends during the summer, all of which may increase the benefits of spending time in a park [36]. Alternatively, more non-residents may be tweeting from parks during the summer, leading to greater within-park sentiment scores. Similar dynamics may be driving the higher happiness benefits on the weekend compared to weekdays, though all days of the week exhibited positive values (See Fig 4). Prior work has shown that people on Twitter are happiest on the weekends and during times of year with more daylight [37]. Nevertheless, our comparisons indicate that a sentiment benefit occurs throughout the day, week, and year, indicating that the effect is not purely driven by temporal patterns. Our hourly comparison indicates that a sentiment benefit occurs during all hours of the day, indicating that the effect is not purely driven by leaving the office. This result is encouraging because some prior studies on nature contact using Twitter analyzed shorter time periods. Future studies should seek methods that can investigate the other temporal aspects of nature contact including the frequency and duration of visits [38].

We acknowledge that studying human behavior using Twitter data involves several potential sources of bias. Active users on Twitter tend to be younger and more affluent than the population at large [39]. Instead of investigating how individual users and demographic sub-groups respond to nature contact, we attempt to estimate the aggregate effect of park visitation on happiness across a city. While our happiness benefit calculation uses same-city tweets as a control, the results may not generalize beyond Twitter users. We only use English language tweets which may limit our ability to generalize to other languages and cultures. We do not control for nearby demographics when assessing the happiness benefit of specific parks. For example, larger parks may be promixal to more affluent neighborhoods or associated with adjacent neighborhood age structure. While this may introduce bias across parks within cities, it should not impact our results comparing the total happiness benefit across cities.

Future directions

Our results, along with those from previous studies, point to several important areas of future research. Future research should continue to explore the relationship between tweet happiness and other factors beyond park investment. While ParkScore® captures a variety of park-quality related metrics, vegetation and biodiversity are salient features of greenspace that significantly impact how people experience their time in nature [40–42].

More localized studies could look at the mental health impact of park-level vegetative cover and biodiversity metrics. Alternatively, similar methods could be applied to compare the mental benefits of nature contact with other experiences such as museum visits or sports games. This could provide insight into the benefit of investing in public goods such as parks for health outcomes relative to alternatives. Similarly, these analyses could isolate the importance of experiencing nature compared to the social and cultural factors that influence sentiment on Twitter.

While we investigated the seasonal variation of in-park happiness, climate and weather have been shown to influence happiness on Twitter as well [43, 44]. Tweets could be binned by some composite of temperature, humidity, and precipitation in order to investigate how weather moderates the association between nature contact and mental well-being [21].

Demographic, socioeconomic, and cultural factors also play a role in how people engage with parks [45]. While identifying such factors on Twitter is challenging and requires ethical consideration, other methodologies can continue to explore how different groups use and benefit from time in parks, to help ensure that the benefits of parks are available to everyone. As the evidence continues to mount on the many different benefits of nature contact, we must ensure park access to quality parks for all urban residents.