Some Implications on Pandemic Green Place Planning and Management: Considering the Patients with Depression

Study area

The study was conducted in Seoul, the capital of South Korea, a densely populated area. Seoul has mountains (e.g., Bukhansan and Dobongsan), waterfront spaces such as the Han River and Jungnangcheon, and large and small parks in the city center, making it a suitable place to explore various green space usage behaviors. Furthermore, there are various types of residents, including office workers, housewives, self-employed, students, and unemployed persons, making it a conducive place to examine and compare the patterns of green space use by various groups (Fig. 1).

Fig. 1

Mountain, river, and park in Seoul.

In this study, we used interviews to explore how natural spaces are planned and managed to treat depression. Previous studies have already shown that visiting natural spaces during the pandemic is effective in treating depression through surveys (Pouso et al., 2021; Dzhambov et al., 2021; Ribeiro et al., 2021, Reid et al., 2022). Therefore, rather than simply creating more green spaces, this study aimed to identify green space planning methods to reduce depression by interviewing the general public and depressed patients about their in-depth use of natural spaces to find out the basis for detailed planning methods. The interviewees were divided into two groups: 6 individuals with depression and 12 individuals without depression (see Table 1). All of them had lived in Seoul for at least 10 years. The participants were selected based on gender, age, neighborhoods in Seoul, and workplaces after posting a public notice throughout the city. The interviews were conducted between April and July 2021. The participants were recruited by publicizing that researchers required individuals who frequently visited mountains, rivers, and parks, i.e., typical green spaces during COVID-19. The six depression patients were recruited from people who had been diagnosed with depression by a hospital and treated for depression. They all suffered from mental illnesses such as depression and insomnia and were receiving psychiatric care. We recorded their reasons, behaviors, and effects of visiting mountains, rivers, and parks during the COVID-19 pandemic.

Table 1

Research participants

Research methodology

Qualitative and quantitative methods have been used equally to analyze each participants' behavior patterns and needs. In qualitative research, the results of focus group interviews are often confirmed through content analysis (Raum et al., 2021) or grounded theory (Banks et al., 2016), according to the framework set by the researcher. However, these methods are limited owing to the researcher's subjectivity. By contrast, surveys, which are representative of quantitative methods, are also used to survey stakeholders on items set by the researcher through prior research; therefore, the precise needs of various stakeholders in the field are not well revealed (Bendtsen et al., 2021). One method that has emerged to compensate for these qualitative and quantitative limitations is text mining—a technique that analyzes messages by dividing them into morphemes and effectively identifies the main content and structure of messages (Park and Leydesdorff, 2013; Shi et al., 2017). In this study, the thoughts, opinions, and feelings of patients with depression and the general public regarding the use of green spaces during the COVID-19 pandemic were analyzed by text mining the transcripts of focus group interviews.

This study used semantic linkage network analysis, a representative text mining technique that identifies the meaning and structure of a message by creating a connection network through the co-occurrence frequency of words used in many social science studies because it can quantify and visualize qualitative data previously identified qualitatively (Lim et al., 2016; Lee, 2019; Lee et al., 2022). The analysis was conducted in the following order. First, we performed a morphological analysis (Bolinger, 1950). Second, among the extracted morphemes, nouns that can have a meaning on their own and high-frequency words are considered to have more critical information power (Luhn, 1958), and the words with the top 10% frequency were selected as the primary keywords. Third, by identifying the frequency of co-occurring words between the contents of a speaker's utterances, we analyzed the connectivity network based on betweenness centrality, which indicates the relationship between words. Based on this, we performed a community analysis with words that were frequently used together. Among them, four segmented words appeared, which were not connected to the main message and were deleted. Fourth, based on the cluster analysis results, we compared the frequency of utterances of patients with depression and the general population using the main content. Fifth, we interpreted the ratio of depression patients to the general population using actual sentences. This study performed morphological analysis and clustering based on betweenness centrality using a Korean text-mining program called Net-miner 4.3 (Fig. 2).

Fig. 2

Research flow chart.

Text-mining

Of the 32 words in the top 10% of words that have meaning alone for the entire script, we excluded four comments that did not have connectivity and selected 28 words. Cluster analysis was performed on these words, and they were divided into two clusters with a best-cut value of 9.790459 (see Fig. 3). In Community 1, there were a lot of words associated with activity. There were many stories about stepping out of the 'home (frequency 126)' and 'exercising (frequency 57)', such as 'walking (frequency 41)' along the 'Han River (frequency 66)' or 'hiking (frequency 21)', breathing fresh 'air (frequency 19)', 'talking (frequency 41)' with 'family (frequency 46)' or playing with 'puppies (frequency 26)', and spending 'time (frequency 11)' engaging in various activities such as exercise and fellowship. In other words, it was a cluster that focused on nature-related activities. Community 2 elicited mostly words related to nature observation. demonstrated many stories of people finding 'ecological (frequency 52)' 'nature (frequency 108)' and experiencing 'seasonal (frequency 18)' changes from the 'mountains (frequency 157)', 'forests (frequency 33)', flowers, and 'trees (frequency 27)'. In other words, it was a group that put the purpose of seeing nature in itself. Table 2 shows the frequency and prominence of each word, and Table 3 summarizes the actual interview transcripts.

Fig. 3

Betweenness community analysis.

Table 2

Betweenness community analysis

Table 3

Interviews by communities

Analyzing the proportions of individuals with and without depression by group showed that those without depression talked more about activities. However, non-depressed people talked about actions and nature equally. Pearson's chi-square test was performed to determine the difference in the purpose of park visits during the coronavirus between depressed patients and the general population, and there was a significant difference with χ²(1) = 33.629, p = .000 (Fig. 4 and Table 4).

Fig. 4

Rate of green space use (activity, nature observation) for people with depression and the general population.

Table 4

Rate of green space use (activity, nature observation) for people with depression and the general population

Socializing and exercising while using greenery considerably helped in alleviating individuals' depression. By contrast, the general public was also interested in the changes in nature, such as the change of seasons and the degradation of nature. These trends suggest that green space planning for people with depression during COVID-19 should enable a range of activities and that alternatives to landscape planting and vandalism management should be made available (Table 5).

Table 5

Interviews of depressed people and non-depressed people

We performed semantic network analysis on the interview transcripts of 6 depressed people and 12 non-depressed people, respectively. The keywords that emerged only from the depressed people were 'alone (frequency 34)' or with 'family (frequency 17) ' and spending 'time (frequency 41)' on 'weekend (frequency 18)' doing various activities such as 'climbing (frequency 21)'. On the other hand, the keywords derived from the non-depressed were 'nature (frequency 92)', 'forest (frequency 33)', 'tree (frequency 27)', 'ecology (frequency 25)', 'season (frequency 18)', and 'flower (frequency 18)', indicating that the ecosystem itself provides a lot of healing (frequency 14). Common keywords were 'mountain', 'Han river', 'exercise', 'excercise', 'friends' and 'walking '. This suggests that while the general public feels a lot of healing just by looking at the natural environment, for depressed patients, active content in green spaces is more effective in reducing depression (Table 6). These results are also evident in the centrality analysis. Depressed people's networks are centered around much more active keywords (Fig. 5). However, in the case of the general population, keywords about nature itself and keywords about activities are connected through 'neighborhood' (Fig. 6).

Table 6

The keywords identified through semantic network analysis for depressed people and non-depressed people

Fig. 5

Betweenness community analysis of depression people's interviews.

Fig. 6

Betweenness community analysis of non-depression people's interview.

Implications for green space planning in the pandemic era

According to the text-mining results of the in-depth interviews for this study, the patients with mental illnesses such as depression or insomnia, and are receiving psychiatric care, were more interested in outdoor activities, such as exercise and socialization, than the general population. Prior researches have demonstrated that COVID-19 has re-inforced the importance of green spaces (Beckmann-Wübbelt et al., 2021; Soga et al., 2021; Jo et al., 2022). Furthermore, studies have highlighted the need for increasing accessibility (Yap et al., 2022) and for socializing in green spaces (Grima et al., 2020). This study is significant because it identifies which parts of green spaces are more necessary for the patients with mental illness. People with depression tend to actively exercise and socialize, since these activities alleviate and treat depression. Exercise has also been shown to help treat depression (Carek et al., 2011; Cooney et al., 2013). This suggests that exercise and socializing with a companion are essential to alleviate depression, rather simply observing nature when using green spaces (Berman et al. 2012). This active use of green spaces by people with depression highlights the need for green space planning to accommodate several activities, including training (Fullana et al., 2020) and socializing (Anderson et al., 2022). Unlike previous studies claiming that green space effectively manages depression during the pandemic, this study shows that green space policies emphasizing exercise and socializing are needed to manage depression.

For people with depression, an accessible neighborhood green space is critical for treatment. Therefore, there is a need to identify more resources for the future development of exercise and socialization-related green spaces near residential areas and to plan for their accessibility (Yap et al., 2022). In addition, ensuring adequate resting facilities and shading (Grima et al., 2020), lighting, and safety for use at different times of the day is crucial (Mayen Huerta and Cafagna, 2021) for people with depression to fulfill the function of socialization when hiking or exercising near residential areas. Moreover, the general public is also interested in the natural ecosystem itself; therefore, policies to plant landscape crops to enhance landscape effects are also necessary (Weinbrenner et al., 2021; Xiao et al., 2023). As the natural environment is likely to be damaged or overcrowded during the pandemic, policies that separate spaces for human use from ecological nature by reinforcing decks and rest facilities would also be adequate (Lee and Choi, 2020). In addition, planning more trails and shelters to prevent overcrowding, limiting the number of people entering each space, or creating a reservation system to distribute the number of people may be helpful. These policies are likely to be effective in forests (Beckmann-Wübbelt et al., 2021), large natural parks (Miller-Rushing et al., 2021), and oceans (Jo et al., 2022).