Quantifying emotional differences in urban green spaces extracted from photos on social networking sites: A study of 34 parks in three cities in northern China
Graphical abstract
Introduction
The sustainability of urban green spaces has become one of the most critical issues that city governments must deal with. On the one hand, it has been proved that the strengthening of urban green spaces will improve the quality of urban life (Shekhar and Aryal, 2019). The way they are planned and designed also have a broad impact on people's health (Martínez, 2021; Jackson, 2003), which is fundamentally related to the quality of a city (Hunter et al., 2015; Schüle and Bolte, 2015). Besides, in the context of rapid urbanization, poor design and urban planning could lead to ineffective supply of public goods and services, which may aggravate public health problems such as obesity (Mayne et al., 2015), lack of physical activity (Sallis et al., 2009), depression, etc.. On the other hand, effective urban green space services have been proved to effectively boost positive emotions (Pretty et al., 2007) and suppress negative ones like anger, confusion, depression and anxiety. Specifically, 5 min of short-term and mild activities on green spaces could result in better outcomes (Barton and Pretty, 2010). Besides, green spaces are especially effective to positively affect children’s emotional health (Ward et al., 2016). Therefore, the United Nations initiative "Healthy Cities" has put urban development issues about inequality, public health, and public services as its priority of concerns (World Health Organization, 2016).
In 2020, a sudden COVID-19 pandemic struck all over the world. During the outbreak and pandemic prevention procedures, people's normal lives and work experiences have been affected greatly, and that further affected their physical and psychological states indirectly (Marinthe et al., 2020; Qian and Yahara, 2020). Specifically, happy emotion significantly dropped, while the recognition about sadness soared (Meléndez et al., 2020), especially in those who were quarantined. In the pandemic era, people eagerly demand for urban green spaces, and are more likely to have positive perceptions in green spaces than in other environments. Besides, previous research has shown that negative emotions fluctuate less among the elderly, while they demonstrate more among young people (under 18 years old) (Zhu and Xu, 2020).
By mapping the emotions and connecting them with urban green spaces (Pánek et al., 2017; Shoval et al., 2018b), previous researchers have drawn a commuter route map and marked positive and negative emotions along the way. Green spaces exhibit an attractive natural environment on the road are positively correlated with positive emotions, while being negatively correlated with negative emotions (Snizek et al., 2013). However, recent studies have explored the benefits from different types of green spaces for specific health, and the results demonstrated that various types of green spaces may have differed effects on moods or health (De Vries et al., 2013). Besides, they have proposed that green spaces should not be seen as a "simply green area,", but should be considered along with their types (Akpinar et al., 2016). The quality, distribution and size of green spaces are some of the decisive factors for urban planning, design and decision-making.
Apparently, emotional differences between genders exist, which has been indicated in some meta-analytical studies. These studies have confirmed that females are advantageous in terms of emotions from infancy to adolescence (McClure, 2000). Females tend to use loving words to describe themselves, while males tend to use confident ones (Ottoni et al., 2013; LaFrance et al., 2003). Emotions seem very complicated, but psychologists have found that most of them can be simplified into a two-dimensional model of emotional valence and arousal (Russell and Snodgrass, 1987). Emotional valence describes the degree of preference, while emotional arousal describes the degree of excitement.
Most past studies used self-report questionnaires and body sensor data to measure human emotions (Niedenthal et al., 2018; Mizna et al., 2013; Silk et al., 2011). However, as facial expression recognition technology (FER) has been growing fast, some studies have utilized social media geotagging images to automatically infer the user's emotions, which are then used as a place attribute to understand the interaction between humans and the environment (Kang et al., 2017; Singh et al., 2017a). For example, a map of the happiness of the world and cities (Kang et al., 2018; Svoray et al., 2018; Kang et al., 2019) can be created from data about emotional security (Pánek et al., 2017). Emotion mapping platform has great potential for urban planners (Pánek and Benediktsson, 2017). But in recent years, big data and especially social media data have helped to enhance the research potential of human behavior and urban place perception (Hansen et al., 2012; Yang et al., 2017; Wood et al., 2013). People frequently use social media platforms, such as Facebook, Twitter, Instagram, and Flickr, to publish and share their views or communicate with each other (Dwyer et al., 2007; Kankanhalli et al., 2005), so that we can understand the urban environment and reveal the hidden characteristics of urban areas (Frias-Martinez and Frias-Martinez, 2014; Kelley, 2013; Lee et al., 2013; Hu et al., 2015). For example, Twitter sentiment analysis is very useful as an important data source for urban planning (Chakraverty et al., 2015; Chapman et al., 2018; Kagan et al., 2015). Combined with the emotion calculation of GIS, we can measure the emotion of the relationship between humans and the environment (Huang et al., 2020a). In addition, social media platforms provide a new opportunity for researchers to understand the city and its social environment, as well as to investigate and study human behaviors from them (Singh et al., 2017b; Liu et al., 2015), measuring societal happiness (Abdullah et al., 2015).
As the most influential social media platform in China and even in the world, Sina Weibo (https://weibo.com/) has established itself as an effective platform for public interaction, communication and collaboration. It has been integrated deeper and deeper in researches about urban green spaces, such as the demand (Zhu and Xu, 2020) and emotional bias (Huang et al., 2020b). Such integration suggests data from Sina Weibo with superior metropolitan geographical locations can be used as an alternative information source not only for urban planners, but also for individuals in business, leisure, and residential areas. Therefore, our purpose is to extract emotions from images with geographic information tags on social media (Sina Weibo), so as to quantify the emotions at urban green spaces in their locations. In this way, the impact of green space on emotion could be better understood, further urging urban planners to promote residents' sense of happiness and health. Specifically, this research focused on the quantification of the emotions carried by green spaces, trying to explore the relationship between them at a city scale.
In this work we ask the following research questions:
Research question 1 (RQ1): Are there differences in the probability of emotions between genders and among ages in urban green spaces?
Research question 2 (RQ2): Are the characteristics of urban green spaces, such as their sizes, normalised vegetation index (NDVI), transportation characters, and types, related to the reported emotions within?
Research question 3 (RQ3): Do the cities present any differences in the emotions in urban green spaces?
Section snippets
Scope of study
34 Green park spaces in the northern provincial capital cities of Harbin, Changchun, and Shenyang were selected as the places of interest (as shown in Fig. 1). The sample size of microblog data is relatively large as these data are derived from provincial capital cities. Additionally, based on China's "urban green space planning standard" (Wang and Wang, 2019), relevant urban park indicators are selected, including the area scale and normalised vegetation index (NDVI) of urban green space to
Results
The high accuracy of face + + API output in face detection is verified in previous research (LaFrance et al., 2003). Therefore, it is reliable for estimations of EC, age, and gender reliability and validity from the data set (Bakhshi et al., 2014).
Conclusion
In this study, the relationship between urban green spaces and emotions is explored by utilizing big data on the Internet. Information about people’s facial expression was extracted and quantified from social media photos, and these emotional results from public perception were further applied for urban green space design.
Results from statistical analyses showed that a higher NDVI for one green space would lead to greater probability and intensity of emotion expression in it. In addition,
The geography of facial expressions
The emotional relationship between human beings and the urban environment has always been the concern of urban scholars. The era of big data has connected emotions and spaces closer by techniques such as real-time measurement (Svoray et al., 2018) and dynamic physiological sensing, which can be used to empirically study the complex interaction between emotions and spaces (Shoval et al., 2018a).
In this paper, FACE++ was applied to identify seven emotions, and public emotions in green spaces were
Funding
This research was funded by The Opening Fund of Key Laboratory of Interactive Media Design and Equipment Service Innovation, Ministry of Culture and Tourism (Project Number 202010), Heilongjiang art science planning project (Grant number 2020A006).
Author contributions
Conceptualization, X.Z. and B.Z.; methodology, X.Z., M.G., and B.Z.; validation, X.Z., M.G., and B.Z.; formal analysis, M.G. and R.Z.; investigation, M.G. and R.Z.; resources, X.Z. and M.G.; data curation, M.G., X.Z., and R.Z.; writing—original draft preparation, M.G. and R.Z.; writing—review and editing,X.Z. and B.Z.; visualization, X.Z., M.G., and R.Z.; project administration, X.Z. and B.Z.; funding acquisition, X.Z.and B.Z. All authors have read and agreed to the published version of the
Declaration of Competing Interest
The authors report no declarations of interest.
Acknowledgments
We would especially like to thank the students from the school of computer science and technology of Harbin Institute of Technology to help data processing.
We thank Elsevier (https://cn.webshop.elsevier.com/) for its linguistic assistance during the preparation of this manuscript.
References (73)
- et al.
Stress rehabilitation through garden therapy: the garden as a place in the recovery from stress
Urban For. Urban Green.
(2013) - et al.
Does green space matter? Exploring relationships between green space type and health indicators
Urban For. Urban Green.
(2016) - et al.
Emotional maps based on social networks data to analyze cities emotional structure and measure their emotional similarity
Cities
(2019) - et al.
Streetscape greenery and health: stress, social cohesion and physical activity as mediators
Soc. Sci. Med.
(2013) The social production of ecosystem services: a framework for studying environmental justice and ecological complexity in urbanized landscapes
Landsc. Urban Plan.
(2013)- et al.
Spectral clustering for sensing urban land use using Twitter activity
Eng. Appl. Artif. Intell.
(2014) - et al.
Residential surrounding greenness and self-reported symptoms of anxiety and depression in adolescents
Environ. Res.
(2021) - et al.
Extracting and understanding urban areas of interest using geotagged photos
Comput. Environ. Urban Syst.
(2015) - et al.
Quantifying the bias in place emotion extracted from photos on social networking sites: a case study on a university campus
Cities
(2020) - et al.
Quantifying the bias in place emotion extracted from photos on social networking sites: a case study on a university campus
Cities
(2020)
The impact of interventions to promote physical activity in urban green space: a systematic review and recommendations for future research
Soc. Sci. Med.
Passion for outdoor activities and environmental behaviors: a look at emotions related to passionate activities
J. Environ. Psychol.
Benefits and well-being perceived by people visiting green spaces in periods of heat stress
Urban For. Urban Green.
Health differences in an unequal city
Cities
Emotion recognition changes in a confinement situation due to COVID-19
J. Environ. Psychol.
Emotional mapping and its participatory potential: opinions about cycling conditions in Reykjavík, Iceland
Cities
Neighborhood environments and physical activity among adults in 11 countries
Am. J. Prev. Med.
Role of geospatial technology in understanding urban green space of Kalaburagi city for sustainable planning
Urban For. Urban Green.
Tracking technologies and urban analysis: adding the emotional dimension
Cities
Daily emotional dynamics in depressed youth: a cell phone ecological momentary assessment study
J. Exp. Child Psychol.
Mapping bicyclists’ experiences in Copenhagen
J. Transp. Geogr.
Demonstrating the effect of exposure to nature on happy facial expressions via Flickr data: advantages of non-intrusive social network data analyses and geoinformatics methodologies
J. Environ. Psychol.
The impact of children’s exposure to greenspace on physical activity, cognitive development, emotional wellbeing, and ability to appraise risk
Health Place
A novel method for human age group classification based on Correlation Fractal Dimension of facial edges
J. King Saud Univ. - Comput. Inf. Sci.
Collective smile: measuring societal happiness from geolocated images
Faces engage us: photos with faces attract more likes and comments on instagram
Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. Toronto Ontario, 26 April 2014–1 May 2014
Robust sentiment detection on Twitter from biased and noisy data
What is the best dose of nature and green exercise for improving mental health? A multi-study analysis
Environ. Sci. Technol.
Emotion–location mapping and analysis using twitter
J. Inf. Knowl. Manage.
Investigating the emotional responses of individuals to urban green space using twitter data: a critical comparison of three different methods of sentiment analysis
Urban Plan.
Show your best self(ie): an exploratory study on selfie-related motivations and behavior in emerging adulthood
Telemat. Inform.
A taxonomy of quality assessment methods for volunteered and crowdsourced geographic information
Trans. GIS
Comment on “100% accuracy in automatic face recognition”
Science
Trust and privacy concern within social networking sites: a comparison of Facebook and MySpace
An argument for basic emotions
Cogn. Emot.
Managing impressions online: self-presentation processes in the online dating environment
J. Comput. Commun.
Cited by (42)
The inequity of urban green space availability between urban villages and residential quarters: An empirical study in Shenzhen, China
2024, Journal of Cleaner ProductionPublic emotions and visual perception of the East Coast Park in Singapore: A deep learning method using social media data
2024, Urban Forestry and Urban GreeningBiodiversity and human health: A scoping review and examples of underrepresented linkages
2024, Environmental ResearchUser-generated content affects urban park use: Analysis of direct and moderating effects
2023, Urban Forestry and Urban Greening