Digital Geography 5 (2023) 100065 Contents lists available at ScienceDirect Digital Geography and Society journal homepage: www.sciencedirect.com/journal/digital-geography-and-society Plural relational green space values for whom, when, and where? – A social media approach Fulvia Calcagni a, b, James J.T. Connolly c, Johannes Langemeyer a, d, * a Department of Geography, Humboldt Universität zu Berlin, Germany b Department of Natural Resources and Environmental Management, University of Haifa, Israel c University of British Columbia, School of Community and Regional Planning, Vancouver, Canada d Institute of Environmental Science and Technology (ICTA), Universitat Autònoma de Barcelona (UAB), Edifici Z (ICTA-ICP), Carrer de les Columnes s/n, Campus de la UAB, Cerdanyola del Vallès 08193, Spain A R T I C L E I N F O A B S T R A C T Keywords: The values people ascribe to their interactions with and within the environment are essential to inform justice Plural relational values and sustainability transformations. The development of many of these values unfolds through enjoying so-called Social media metadata cultural ecosystem services (CES) such as outdoor recreation, landscape aesthetics or environmental education. A Geo-temporal-demographic cross-analysis growing body of literature is improving the assessment of the multiple ways that people value human and non- Justice and sustainability transformations human relations arising when enjoying CES. Yet, the geo-temporal-demographic patterns of values distribution and the lessons that can be derived are to be consistently analysed within this relational framework. Building on a visual and textual content analysis of social media (SM) data geotagged in a peri-urban park of Barcelona, Spain, this research explores the potential of analysing the associated metadata (such as geotag, timestamp and social media users’ demographics – i.e., performed gender and residency) in order to develop a better under­ standing of the linkages between people’s values and the situated context of their construction. Our results show trends in relational CES values distribution along and between the analysed spatial, temporal, and demographic dimensions. In particular, despite there being a multiplicity of values revealed across the whole case-study area, to enjoy contemplative CES, such as spiritual or cognitive value, people need to move away from highly frequented areas and prefer specific times of the day, respectively evening or afternoon. Locals show a higher preference to visit the park on weekends compared to non-locals, while women-performing users show a significantly higher drop in their CES benefits uptake compared to men-performing users at night. In addition to providing novel and fine-grained information for transformative practices toward justice and sustainability, this study highlights the importance of complementing CES studies employing SM with metadata analysis to improve our understanding of the relationship between the real and the more-than-real. 1. Introduction material and non-material benefits that they value and, in turn, contribute to produce (Dickinson & Hobbs, 2017). As a result, ES is an Understanding people’s social-ecological interactions and related established framework for integrating numerous social and environ­ benefits uptake is pivotal to the development of pathways to sustainable mental factors into decision-making (European Commission, 2020; futures that accounts for the role of underlying value structures in Plieninger, Dijks, Oteros-Rozas, & Bieling, 2013) and is increasingly shaping outcomes (Arias-Arévalo, Gómez-Baggethun, Martín-López, & gaining importance in fostering just and sustainable planning agendas Pérez-Rincón, 2018; Guo, Zhang, & Li, 2010; Ilieva & McPhearson, (Ernstson, 2013; IPBES, 2018; Langemeyer & Connolly, 2020). This role 2018; Kremer et al., 2016; MEA, 2005; Plieninger et al., 2015; Riechers, for ES analysis highlights the importance of measuring differential ac­ Barkmann, & Tscharntke, 2016; Stålhammar, 2021). This notion is well- cess to and uptake of ES benefits, especially in ever-denser and developed within the ecosystem service (ES) scholarship, a now long­ increasingly ethnically and linguistically diverse metropolitan areas standing effort to comprehensively account for the biophysical struc­ (Amorim-Maia et al., 2020; Andersson et al., 2019; Baró, Langemeyer, tures, processes, and functions from which people derive multiple Łaszkiewicz, & Kabisch, 2021; Calderón-Argelich et al., 2021; Danford, * Corresponding author at: Institute of Environmental Science and Technology, Autonoma University of Barcelona E-mail addresses: fulvia.calcagni@gmail.com (F. Calcagni), james.connolly@ubc.ca (J.J.T. Connolly), johannes.langemeyer@uab.cat (J. Langemeyer). https://doi.org/10.1016/j.diggeo.2023.100065 Received 19 April 2022; Received in revised form 23 June 2023; Accepted 4 July 2023 Available online 7 August 2023 2666-3783/© 2023 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Strohbach, Ryan, Nicolson, & Warren, 2014; Hamstead et al., 2018). recently pointed at the multiplicity of ways and plurality of sub­ Yet, accounting for ES benefits that people actually experience and value jectivities from which individuals value social-ecological interactions, remains an open challenge. These benefits are possible expressions of raising the need to include a wider range of socio-cultural values into ES the situated and subjective “view from the body” advocated by Har­ accounting beyond only economic terms (Andersson, Tengö, McPhear­ away, 1988 (pg. 589), elements of a dynamic and relational processes son, & Kremer, 2015; Arias-Arévalo et al., 2018; Chan et al., 2012; Chan, through which they construct their sense of belonging in space (Bee­ Satterfield, & Goldstein, 2012; Díaz et al., 2015; Dickinson & Hobbs, beejaun, 2017), and descriptive of current states of social values key to 2017; Fischer & Eastwood, 2016). informing pathways toward sustainable transformations (Stålhammar, 2021). This is especially the case for those benefits ascribed to so-called 1.1.1. Values for sustainability cultural ecosystem services (CES), which are less tangible and objective ES functions become services when people, through their system of (e.g., landscape aesthetics, cultural heritage, or simply the appreciation situated knowledge, beliefs and practices, value the benefits they pro­ for the existence of a natural resource). vide (Calderón-Argelich et al., 2021; Chan, Satterfield, & Goldstein, In addition, measuring uptake and access to the specific range of CES 2012; Dickinson & Hobbs, 2017). In other words, ES benefits uptake is benefits that people value because contributing to individual wellbeing always dynamically produced and negotiated in space (Beebeejaun, or to collective flourishing can be seen as an essential “technology of 2017), as well as heavily mediated by human values and the cultural, participation” needed to increase public acceptance, equity and reflex­ historical, spatial and political context in which humans live and build ivity in planning and transformative measures (Angradi, Launspach, & their capabilities and identities (Fischer & Eastwood, 2016; Fish et al., Debbout, 2018; Ilieva & McPhearson, 2018; Pascual et al., 2017; Tadaki, 2016; Harvey, 1996). In line with this thinking, we assume that cultural Sinner, & Chan, 2017). Therefore, for the purpose of value-informed ecosystem services (CES) are not just a subset of ES, as commonly decision-making toward justice and sustainability, scholars from assumed (MEA, 2005). Rather, CES provide irreplaceable access to and different fields (e.g., environmental science, digital critical geography, operationalisation of the non-material, intangible, and incommensu­ feminist digital nature and urbanism) call for consideration of the rable benefits arising from social-ecological interactions and, in turn, multiple and intersecting factors influencing the ways in which we value from ES functions (sensu Andersson et al., 2015). Values ascribed to CES ecosystems, and encourage the measurement of the spatial, temporal benefits are, thus, regarded as essential and pervasive inputs to, both, and demographic variability in the benefits gained from CES (Beebee­ the realization of those same benefits and the valuation and consequent jaun, 2017; Blicharska et al., 2017). From this perspective, a special production of the underlying ES (Dickinson & Hobbs, 2017; Fischer & stress is placed on acknowledging the plural – and often conflicting – Eastwood, 2016; Langemeyer & Connolly, 2020), pointing toward an valuation languages and priorities we associate with social-ecological important set of motivations that serve as a driver for ES conservation interactions (Arias-Arévalo et al., 2018; Himes & Muraca, 2018; Ja­ (Andersson et al., 2015; Dickinson & Hobbs, 2017; Stålhammar, 2021). cobs et al., 2018; Pascual et al., 2017; Small, Munday, & Durance, 2017; Mapping, monitoring and assessing CES has proven critical to define Tadaki et al., 2017). sustainable use of natural resources and restoration targets (European Yet, CES values assessments are straining to meet emerging demands Commission, 2020) as well as to perform impact and strategic assess­ for more plural and relational information, in part because they are ments (namely Environmental Impact Assessment and Strategic Envi­ mostly conducted through traditional data collection methods, such as ronmental Assessment) (Geneletti, 2016). Including CES values in interviews, surveys, focus groups or participatory mapping (Arias- decision-making is, thus, essential to foster sustainability and resil­ Arévalo et al., 2018; Dickinson & Hobbs, 2017). While these approaches ience in transformative practices. do offer important insights, they are time- and resource-intensive, and thus heavily conditioned by their spatial, temporal, and demographic 1.1.2. Values for justice scale and sensitivity (Milcu, Hanspach, Abson, & Fischer, 2013; Riechers The way to incorporate distributive, procedural, and recognition et al., 2016). Traditional data collection approaches also express sys­ justice into decision-making is starting to be explored. For example, tematic biases as a result of research design and operation (Derungs & building on precedents from non-ES studies (see Currie, Lackova, & Purves, 2016; Lopez, Magliocca, & Crooks, 2019; Pastur, Peri, Lencinas, Dinnie, 2016; Krange & Skogen, 2007), some authors find that socio­ Garcı-Llorente, & Martìn-Lopez, 2016; Tadaki et al., 2017), likely only economic factors influence the uptake of CES. Fortnam et al. (2019) look reflecting a partial memory of the stated values, more conditioned by at gender differences in CES perceptions; Martinez-Harms et al. (2018) normative understanding of values, and falling short in uncovering the reveal differential socio-economic impact on visiting protected areas; benefits experienced at the actual site and moment of realization. With and Plieninger et al. (2013) stratify their analysis of CES perceptions by this study, by employing social media data and metadata, we aim to socioeconomics. assess the fine-grained spatial, temporal and demographic variation of a A full understanding of the range of benefits, and multiple value broad scope of revealed relational CES values relevant to environmental systems for defining those benefits, associated with CES is an important justice and sustainability in decision-making. We do so by, first, defining element of social-ecological justice (Calderón-Argelich et al., 2021; the values that we specifically focus on; then outlining the data and Grossmann, Connolly, Mattioli, & Nitschke, 2022). This is the case methods chosen to assess them; and, finally, describing the results from because, firstly, the uptake of CES benefits depends more strictly on the a case study application that relates the findings to possible insights for actual occurrence of and access to social-ecological interactions than is planning. the case for other types of ES (e.g. climate regulation, carbon seques­ tration, rainwater runoff, etc.), which can be supplied by distant eco­ 1.1. Notions of value in the ES framework systems (Martinez-Harms et al., 2018). Secondly, because values ascribed to CES benefits have recently been defined as inherently rela­ The value ascribed to ES is understood as the relative importance tional (Chan et al., 2016; Chan, Gould, & Pascual, 2018). individuals assign, either individually or through group negotiations, to the several, interdependent benefits the ecosystem provides (Chan, 1.1.3. Relational CES values for transformations Satterfield, & Goldstein, 2012). Reflecting a dichotomous understanding Relational values are described as socio-cultural, anthropocentric, of society and nature as separate and independent ontological domains, yet non-instrumental values rooted in social-environmental relations and an epistemological distinction between the subject and the object of that people seek to maintain and find desirable for individual and col­ valuation drawn from market theories rooted in political economic lective flourishing (Himes & Muraca, 2018), linking them with core thought of the eighteenth century (Harvey, 1996), ES values were social values such as justice, care, and reciprocity (see Chan et al., 2016). initially translated into mainly economic terms. However, scholars have Relational values ascribed to CES (hereafter referred to as relational CES 2 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 values for the sake of emphasizing this link) point toward a nuanced relations with people and nature that they benefit from (thus value and middle ground of human and non-human relations (Dickinson & Hobbs, benefit are tightly interlocked and both terms are used throughout the 2017; Fish, Church, & Winter, 2016; Muraca, 2016; Panelli, 2010; article). These data, thus, as crowdsourced and unsolicited “view from Stålhammar & Thorén, 2019) and are at the centre of any dynamic somewhere”, can substitute the “view from nowhere” provided by process of space and time construction, negotiation and signification authoritative and official data – to which access and availability for (Beebeejaun, 2017; Harvey, 1996; Massey, 1994). In addition, besides research is also decreasing (Shelton, 2023) – away from the objectivity being important components of human wellbeing, relational values are guise and in the sake for different, even if yet partial, categories and also thought to enable plural and subjective languages of valuation, so methods for assessing spatial and social truth and justice (Shelton, that their consideration in transformative practices improves inclusivity 2022). Further, this expression of value can be measured in terms of a and recognition of different people’s claims and needs (Chan et al., continual spatial and temporal pattern due to the ongoing log formed by 2018; Díaz et al., 2015; Himes & Muraca, 2018; Pascual et al., 2017) social media data (Ghermandi & Sinclair, 2019). Thus, by revealing allowing to expand the conceptualization of justice toward a less top- preferences expressed in different ‘languages’ and formats (e.g. through down and more crowdsourced form. pictures, texts, “likes”, etc.) that are freely shared and unbiased by Relational values’ space- and time-dependency, collective negotia­ researcher intervention, and including metadata such as geotags, time tion, and expressions of people’s plural conception of a meaningful stamps and demographics, SM offer promising opportunities for rela­ community life within nature, make their assessment well-suited to tional CES values assessment to explore a broader epistemological plu­ meeting the complex demands of value-informed transformative prac­ rality (i.e. recognition justice) (Armstrong, Derrien, & Schaefer-Tibbett, tices devoted to justice and sustainability (Blicharska et al., 2017; Himes 2021; Barry, 2014; Calcagni et al., 2019; Ernstson, 2013; Leszczynski & & Muraca, 2018; Pascual et al., 2017; Tadaki et al., 2017), being not Elwood, 2015; Martinez-Alier, Kallis, Veuthey, Walter, & Temper, 2010) only key drivers for environmental stewardship and for advancing sus­ and inquire into the differential patterns (i.e. distributional justice) tainable and just transformative agendas as we know them, but also a (Ilieva & McPhearson, 2018; Martinez-Harms et al., 2018) of benefits marker for a different pathway of development for these agendas. access and uptake. In addition, building on the branch of feminist urbanism scholarships SM metadata are increasingly available at an unprecedented rate and focusing on the right to the everyday life (Beebeejaun, 2017), we argue scale (Ilieva & McPhearson, 2018), allowing for (near) real-time, high that the analysis of relational CES values engendered over embodied resolution, spatially explicit and, sometimes, global analyses (Huang, tactics of benefits realization at the fine-scale of the “everyday” could Gartner, & Turdean, 2013; Tieskens, Van Zanten, Schulp, & Verburg, provide novel insights into questions of social-ecological justice and 2018; van Zanten et al., 2016). This data helps explain the values at the sustainability transformations, especially in the urban context. moment and site of their realization (Ghermandi & Sinclair, 2019). However, despite a steadily growing body of literature, the attempts Besides CES studies, geodemographic analyses of SM data have proved to empirically assess and interpret the meaning of relational values useful in assessing how demographic factors relate to human mobility remain limited. Most studies use qualitative research approaches alone patterns (Luo, Cao, Mulligan, & Li, 2016), activity spaces (Shelton, or mixed methodologies, through open-ended questions with a non- Poorthuis, & Zook, 2015), neighbourhoods’ available facilities (Quercia predefined analytical framework (Arias-Arévalo et al., 2018; Jacobs & Saez, 2014), as well as to landscape appreciation and accessibility et al., 2018; Schulz & Martin-Ortega, 2018). In addition, recent studies (Martinez-Harms et al., 2018). Among the wealth of empirical studies have highlighted research gaps in addressing questions of justice within using SM data for CES assessment, the majority use the attached geotag – ES assessments, highlighting the particular need to focus on relational or location-related tags (Alieva et al., 2021; Bernetti, Chirici, & Sac­ values, conduct analyses across spatial scales and along inter- and intra- chelli, 2019; Gosal & Ziv, 2020) – for inferring the spatial distribution of generational temporal scales, include gender and intersectional per­ CES values (Calcagni et al., 2019; Hamstead et al., 2018; Havinga, spectives, and combine the different dimensions of distributive, proce­ Bogaart, Hein, & Tuia, 2020; Wood, Guerry, Silver, & Lacayo, 2013) and dural and recognition justice (Calderón-Argelich et al., 2021; some others also use the timestamp (Barros, Moya-Gómez, & García- Langemeyer & Connolly, 2020). Palomares, 2019; Crampton et al., 2013; Gosal, Geijzendorffer, Václavík, Poulin, & Ziv, 2019; Mancini, Coghill, & Lusseau, 2018; 1.2. The potential of using social media metadata Shelton et al., 2015; Sonter, Watson, Wood, & Ricketts, 2016; Walden- Schreiner, Rossi, Barros, Pickering, & Leung, 2018) and show the val­ Social media (SM) is emerging as a novel approach to eliciting a fine- idity of this data in approximating visitation rates and temporal patterns grained assessment of multiple relational CES benefits in the place and at by comparing them to official statistics (Tenkanen et al., 2017). Finally, the moment of their realization (Calcagni, Amorim Maia, Connolly, & some studies associate SM data to demographic information at several Langemeyer, 2019; Calcagni, Nogué Batallé, Baró, & Langemeyer, 2022; scales (Hamstead et al., 2018), some cluster the users based on their Ghermandi et al., 2023; Ruiz-Frau et al., 2020). Unsolicited pictures, pictures content (Gosal, Geijzendorffer, Václavík, Poulin and Ziv, 2019), texts, and videos portraying social-ecological interactions shared on SM, some others extract user demographics through deducing them from joined with the digital network’s inherent “observer effect” (Ghermandi available information (e.g. language spoken) (de Juan, Ospina-Álvarez, & Sinclair, 2019; Lazer, Hargittai, Freelon, Gonzalez-bailon, & Munger, Villasante, & Ruiz-Frau, 2021) or through algorithms which calculate 2021) and presentational (Meikle, 2016) purpose, reveal the their most frequented locations and either correlate the result to census relationship-laden nature of the values expressed therein (Leszczynski, data (Martinez-Harms et al., 2018) or validate it through a survey 2019). The act of sharing is the central aspect that makes SM content (Lenormand et al., 2018). Yet, only a few CES studies have started to meaningful and material to people (Leszczynski, 2019). It generates a explore metadata other than the geotag, such as place of residence link to what they find, and believe observers to find, desirable or just. (Bandara & Bandara, 2019; Cao, Wang, Su, & Kang, 2022; Ghermandi, This reading of the digital as a place for novel encounters between Camacho-valdez, & Trejo-espinosa, 2020; Hamstead et al., 2018; Huai, humans and non-humans, fertile for caring and cultivating shared eco- Chen, Liu, Canters, & Van de Voorde, 2022) and/or gender included in centric goals for collective flourishing, is also gaining momentum or inferable from the user profile (Angradi et al., 2018; Ding, Yang, & within the emerging digital ecology scholarship (Büscher, Koot, & Luo, 2021), despite the explicit recognition of this methodological gap in Nelson, 2017). the context of social-ecological sustainability research to inform Based on the assumption that people use SM platforms to express and decision-making (Alieva et al., 2021; Blicharska et al., 2017; Clemente negotiate individual opinions on what they value (Amorim-Maia et al., et al., 2019; Gliozzo, Pettorelli, & Haklay, 2016; Ilieva & McPhearson, 2020; Calcagni et al., 2019), SM data arguably expose “situated” rela­ 2018; Oteros-Rozas, Martín-López, Fagerholm, Bieling, & Plieninger, tional CES values through people’s outward expression of ongoing 2018; Ruiz-Frau et al., 2020; Tadaki et al., 2017). 3 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Table 1 Relational CES values coded in (Calcagni et al., 2022), respective description, examples of annotation and number of corresponding entries. Coding Description Examples of the annotation for both textual and visual content Nr. of entries categories (% of the total) The engagement, use or enjoyment of the biophysical characteristics or Physical Bikes, walking or running gear, horsing facilities, biking, walking, qualities of species or ecosystems in ways that require physical and 476 (28%) recreation riding or running activity cognitive effort. It denotes an active involvement with nature. The engagement or enjoyment of the biophysical characteristics or Observational or passive activities: relaxing, observing, thinking, Experiential qualities of species or ecosystems through passive or observational taking artistic pictures of nature (close-up pictures of species were 642 (37,9%) recreation interactions. It is an experiential use of plants, animals, and landscapes. categorized also as Existence value) It denotes a passive engagement. Characteristics or features of living systems that have an existential Close-up pictures or common names of (animal or vegetation) Existence value value. The things in nature that people seek to preserve because of their 582 (34,4%) species non-utilitarian qualities and that want to be kept for future generations. Intellectual interactions with the natural environment that foment Knowledge transmission activities, studying outdoors, taking Cognitive value scientific investigation, the creation of traditional ecological knowledge, 183 (10,8%) samples education or training. It is the in-situ research and study of nature. Natural cultural Intellectual interactions with the natural environment that help people Natural landmarks with symbolic significance, iconic 858 (50,7%) heritage identify with the history or culture of where they live and come from. Intellectual interactions with the natural environment that enable Landscape aesthetic experiences. It is the appreciation of the inherent beauty of the Pictures with a wide landscape framing. Tags related to landscape 665 (39,3%) aesthetics biophysical characteristics or qualities of species or ecosystems. It is the scenic value. beauty of nature. Expressions of attachment to religious values, relating to or Elements of living systems that have sacred or religious meaning. Things Spiritual value involving spiritualism, denoting a spiritual state or relating to 21 (1,2%) in nature that have spiritual importance for people. sacred matters. Pictures and text capturing social interactions in the engagement or Images of or tags related to people sharing time and activities in Social relations* enjoyment of the biophysical characteristics or qualities of species or 237 (14%) nature. ecosystems. Intellectual interactions with the built environment that help people Built cultural identify with the history and culture of where they live and come from. Built human infrastructure and landmarks 983 (58,1%) heritage* Cultural heritage or historical knowledge. * Note: Coding categories not included in the CICES reference system (v. 5.1) (Haines-Young & Potschin, 2018). Accounting for the composition of SM data producers and content is 2. Data and methods a necessary step for relating the fine-scale patterns of social-ecological interactions that this specific kind of data reveal to the subjectivities 2.1. Study case and data behind them (Arts, Fischer, Duckett, & Wal, 2021; Elwood & Leszc­ zynski, 2018). In addition, apart from leveraging SM data for investi­ This study is based on the case of Collserola, a large peri-urban park gating dynamics of social-ecological justice on the ground, in this study located north of the highly densified city of Barcelona, Spain. Covering we also look at SM as subject of study. Drawing on considerations an area of almost 8.300 ha, Collserola is the central greenspace in the attempting to overcome the dichotomic understanding of digital spaces - Barcelona Metropolitan Area. Declared a protected Natural Park in and inherent dynamics - as “non-real” and opposed to the “real” of the 2010, Collserola is home to numerous environmental education and non-digital world, we adopt the more-than-real alternative conceptual­ recreation activities. It includes two natural reserves, La Font Groga and ization (McLean, 2020) and recognize its role in co-constructing social- La Rierada-Can Balasc, and it has been the object of several studies and ecological interactions along dimensions of gender, age, race, class, etc. management plans. Such efforts have tried to monitor and manage the In so doing, we respond to a recent call to foster a dialogue between the visitation flow in order to maximise the benefits uptake for park users scholarships of digital geography, digital natures and feminist political and, at the same time, minimise their impact on both the vital ES ben­ ecology, differently engaging with the interactions between the digital, efits supply and the potential crowd effect in specific areas, such as the environmental and social domains (Nelson, Hawkins, & Govia, 2022). area surrounding the Tibidabo mountain, where visitor crowding may The use of these data as both object and subject of study, as fostered in prevent an enjoyable use of the park (Comissió institucional del Pla the digital geography literature, allows for transformative practices to especial de Collserola, 2019; Farías-Torbidoni & Morera Carbonell, unfold in the assessment of both situated and plural relational CES 2020). values ascribed to everyday experiences and their geo-temporal- This study builds on the data and outcomes of a preceding analysis demographic distribution, and in understanding the role of the more- conducted by (Calcagni et al., 2022) in the same case study area and than-real in the co-construction of the real, by challenging official using social media data to assess the multiple relational values attached data positivist biases and embracing the fuzziness of the crowdsourced. to the CES benefits provided by this specific greenspace (see Table 1). Therefore, by combining SM data and metadata, such as the geotag, For said study the authors retrieved a total of 5170 pictures on the timestamp and demographic information, and conducting a “reality photo-sharing platform Flickr using the Application Programming check” with ground data, this study aims to unveil useful insights for just Interface (API). The pictures were uploaded between 2004 and 2017 and and sustainable transformative practices. This is accomplished through geotagged within the boundaries of Collserola. The assessment was (i) fine-scale mapping of the distribution of multiple relational CES performed by three different researchers looking at either or both the values, (ii) evaluating how selected spatial, temporal, and demographic visual and textual content of the retrieved data and coding each item of predictors help explain the distribution of relational CES values, and, data with as many CES as those identified. The non-exclusionary nature finally, (iii) examining qualitative and distributional social inequities in of the assigned codes explains why the percentages of data per category the uptake of CES benefits across space and time. in the last column of Table 1, as well as in the graphs showing relative proportions below, does not sum to 100. The codes were assigned using CICES as a reference system for CES assessment, leveraging the 4 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Table 2 Data source and calculation of the geo-temporal-demographic predictors used in the analysis. Predictors Relation Type Source/calculation Touristic Planar distance Continuous amenities Department of Territory and Sustainability of the Catalan Government (Generalitat de Catalunya) Bus/train Planar distance Continuous stations Spatial Extracted from the topographic map (scale 1:5000), available for download at the website of the Walking paths Planar distance Continuous Cartographic and Geologic Institute of Catalonia (ICGC, http://www.icgc.cat/es/) Calculated performing an Optimized Hot Spot analysis across the full sample of relevant pictures Picture Planar distance Continuous (Aggregation Method: count incident within fishnet polygons; Bounding Polygon: Collserola; Cell Hotspots Size: 400 m) 1/0 Categorical Morning: when pictures “data taken” indicate a time between 06:00 and 12:00 1/0 Categorical Afternoon: when pictures “data taken” indicate a time between 12:00 and 18:00 Time of the day 1/0 Categorical Evening: when pictures “data taken” indicate a time between 18:00 and 24:00 1/0 Categorical Night: when pictures “data taken” indicate a time between 00:00 and 06:00 Weekday: when pictures “data taken” indicate either Mondays, Tuesdays, Wednesdays, Thursdays or Day of the 1/0 Categorical Temporal Fridays week 1/0 Categorical Weekend: when pictures “data taken” indicate either Saturdays or Sundays 1/0 Categorical Summer: Pictures “data taken” between 22/06 and 23/09 1/0 Categorical Autumn: Pictures “data taken” between 24/09 and 21/12 Season 1/0 Categorical Winter: Pictures “data taken” between 22/12 and 20/03 1/0 Categorical Spring: Pictures “data taken” between 21/03 and 21/06 Performed Woman-performing/ Retrieved manually from Flickr user pages based on first names and/or profile pictures (as in Angradi Categorical Demographic gender Man-performing et al., 2018) Residency Local/Non-local Categorical Retrieved manually from Flickr user pages (where given) versatility of this system to adapt the coding to the crowdsourced and at some distance from where they are generated (e.g. aesthetic benefits), passive nature of the data under study (i.e., not dependent on the study the buffer takes into account the CES whose benefits uptake is enabled design but rather on the collective choices of social media users). For by the park even when associated with features outside of its adminis­ instance, we added or merged categories when needed. Coders itera­ trative boundaries (e.g. the Tibidabo mountain). tively checked their agreement through the Cohen’s kappa indicator and In addition, we assume that not only the location but also the time in co-developed a protocol for increasing the consistency and replicability which a picture is taken is an essential characteristic related to the of the analysis. By looking at SM metadata, this study complements the content and the photographer (e.g. location, user’s performed gender, understanding of the multiplicity of relational CES values unveiled in CES valued) that helps to analyse the pictures’ content variety, and (Calcagni et al., 2022) and, in so doing, leverages SM data potential for relative CES benefit uptake over time (Blicharska et al., 2017; Vaz et al., addressing justice and sustainability questions in the assessment of 2020). Therefore, using the pictures’ entry for the field “date taken” relational CES values. provided by the Flickr API, we determined the corresponding time of the day, day of the week and season in which they were taken (approxi­ mating the same date for summer and winter solstices and spring and 2.2. Geo-temporal-demographic analysis of CES uptake autumn equinoxes for every year) (see Table 2). Finally, we accounted for two main demographic characteristics, Building on the same dataset used and analysed in (Calcagni et al., namely gender and residency, which were found to be determinant of 2022), constituted by a total of 1692 relevant pictures, we explored the differential landscape perception and use in previous studies (Plieninger potential of SM metadata by correlating the multiple relational CES et al., 2013). Regarding residency, people declaring that they currently values assessed therein (see Table 1) to the geo-temporal-demographic live within the nine municipalities enclosing Collserola Natural Park variables elicitable from Flickr. (namely, Sant Cugat del Vallès, Molins de Rei, El Papiol, Barcelona, To analyse the spatial distribution of relational CES values, we first Cerdanyola del Vallès, Sant Feliu de Llobregat, Sant Just Desvern, aggregated the data by creating a grid of 124 cells of 1 km2 each through Montcada i Reixac, Esplugues de Llobregat) were coded as ‘local’, all the open-source software QGIS 3.10.10-A Coruña1. Then, we computed others were coded as ‘non-local’. While residency is provided by the the sum of the observations for each relational CES value category per Flickr API, since the 2017’s redesign of the Flickr Profile page, gender grid cell. Subsequently, to understand the correlation between this dis­ ceased to be collected among the personal information and was deleted tribution and a set of chosen spatial predictors (see Table 2 and Fig. 1), from existing profiles, leaving to each user the only option to share it in we used ArcGIS (10.6) to determine the planar distance from the loca­ the description box. Therefore, not being included in the information tion of each sample point to the closest elements of the spatial pre­ retrievable through the API, we assumed gender manually from Flickr dictors. While touristic amenities, bus/train stations and walking paths users’ page (Hamstead et al., 2018; Martinez-Harms et al., 2018; Wood represent the park’s facilities and accessibility factors that have been et al., 2013) by looking at their names and/or profile pictures and widely used in similar studies (Hamstead et al., 2018; Richards & Friess, excluding those ambiguous or unclear. Given the interpretation based 2015; Vaz et al., 2020), pictures hot spots (i.e. spatial clusters of images on users’ appearance online and not on their biological traits, here we posted to Flickr) comprise an innovative factor introduced in this study. refer to gender – not to sex, as approximated elsewhere (Wilkins, 2004) We assumed hotspots function as a proxy for the most frequented places, –, understanding it as a social construct which underlies norms, expec­ interpreted as those where people feel safe, which attract them, or are tations, attitudes and behaviours related to the different combinations of better suited for seeking social approval by sharing them on SM. what is regarded as feminine and masculine (Rogg Korsvik & Rustad, Following evidence from other studies (see Plieninger et al., 2013), 2015). For urban studies in particular, gender assumptions by gender all spatial predictors have been extracted using a 1 km buffer from the presentation are considered a valid methodology for determining policy Collserola boundaries. Acknowledging that CES benefits can be derived implications for public space. This is because, within public spaces, all users are either consciously or subconsciously assuming gender via gender presentation and therefore their interaction with the space or 1 https://qgis.org/en/site/ with others (e.g., fearing dark areas, trusting or aggressive attitudes) are 5 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Fig. 1. Case study area of Collserola park, surrounded by the 9 neighbouring municipalities, including Barcelona. On the left, land cover map; on the right, chosen spatial predictors within 1 km buffer from the park boundaries. dictated as such (Valera & Casakin, 2022). We argue that a similar Wilk’s test for normality. All predictors showed a non-normal distribu­ assumption can be made for social media environments, understood tion and, more precisely, a positively skewed histogram. Consequently, here as virtual spaces of daily life performance and social interaction in we conducted Chi-Square tests of independence between categorical which people negotiate their identities (e.g., through freely choosing variables and Kruskal-Wallis tests between categorical and continuous their username and profile picture) (van Doorn, 2011). variable, as a non-parametric method for the analysis of variance for k ≥ Therefore, while observatory methods in public space rely on gender 2 levels of a factor. Then, in order to calculate pairwise comparisons presentation, we argue that on social media we observe gender perfor­ between group levels with corrections for multiple testing, we used the mance (van Doorn, 2011). However, despite the liberation potential Wilcoxon rank sum test. sometimes ascribed to SM, allowing dissident and non-normative sub­ In addition, we used the Geoda software to detect spatial outliers jectivities to “think/do/be otherwise” (Byron, 2020; Elwood, 2021; through box and scatter plots. Across all predictors, outliers are mainly Pilcher, 2017; Solombrino, 2018), and probably due to the above located in the area surrounding the Tibidabo mountain, the most tour­ mentioned “observer effect” (Ames & Naaman, 2007; Ghermandi and istic and panoramic section in the study area because of the historical Sinclair, 2019) relating people’s behaviour on SM to the perceived amusement park located on its summit. Therefore, since we believed presence of – mostly unknown – others (Meikle, 2016), studies show the that their presence did not depend on poor experimental design but, predominance of heteronormative gender performativity online (van instead, revealed local patterns of CES access, we did not dismiss outliers Doorn, 2011). Thus, for this reason, and for the purpose of our study to from further analyses. Of note, observations for spiritual value had a understand how gender performance (not identity) online weights on relatively low sample size and the results should thus be interpreted as and relates to people’s experience of real and more-than-real public preliminary. spaces, we believe that the binary assumption of SM users’ gender by Finally, we were able to validate SM data through comparison with interpreting the – non-ambiguous – chosen username or appearance on data available from other studies: a survey to Collserola park users profile pictures that they feel comfortable in showing publicly (van conducted between 2017 and 2019 (Farías-Torbidoni & Morera Car­ Doorn, 2011), is acceptable. bonell, 2020), and another conducted by the municipal government of Barcelona in 2016 about gender violence (Adjuntament de Barcelona, 2.3. Descriptive statistics and analysis 2018). In order to check the statistical significance of the relationships be­ 3. Results tween the analysed predictors and the relational CES values’ geo- temporal-demographic distribution, having both categorical (e.g. de­ 3.1. Spatial distribution mographics, time predictors) and continuous (e.g. distance from spatial predictors) variables, we performed a set of Exploratory Data and Spatial Through the geotag, we categorized the spatial distribution of mul­ Data Analyses (EDA and ESDA) with R and Geoda2 (version 1.14) tiple relational values ascribed to CES in the study area (Fig. 2). In respectively to select the adequate descriptive statistical tests. addition to the variety of relational CES values showed by the pie chart, In particular, in R we plotted histograms and performed the Shapiro- the amount of SM content shared in each cell is represented by the corresponding graduated colour. The south-western area is, thus, where most of the SM activity takes place. Interestingly, the amplitude of the 2 See https://geodacenter.github.io/ values spectrum per standard area grid cell appears to be linked to the 6 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 respective amount of SM data: the fewer the SM posts, the lower the with seasonality, with physical recreation (p-value = 0.01088) and social variety of CES benefits uptake. One way of understanding this trend is to relations (p-value = 0.001312) mainly enjoyed during autumn, while see it as reflecting a circumstance where, more people go to areas with a landscape aesthetics (p-value = 0.01931) mainly in spring (similarly to wider variety of CES benefits. Overall, frequent values include built Cao et al., 2022). cultural heritage, natural cultural heritage, physical recreation and landscape aesthetics. 3.2.3. Demographic predictors Finally, we assessed demographic differences in the uptake of CES 3.2. Statistical correlation between geo-temporal-demographic predictors benefits. Out of the 471 unique users in our sample, we determined the and CES uptake performed gender of 433 (92%): 55 (13%) performed as women and 376 (87%) as men, in similar proportions to other studies (see Angradi et al., 3.2.1. Spatial predictors 2018). Two Flickr accounts are shared among a couple and were not The distance that people are willing to walk from the spatial pre­ included in demographic analyses. In relation to the number of photos dictors listed in Table 2 (i.e. touristic amenities, bus/train stations, per performed gender category, out of the total 1617 pictures uploaded walking paths, and picture hotspots) varies, reflecting different spatial by photographers whose performed gender we could assign, 92% (1482) patterns associated with the various CES benefits that people pursue. were taken by men-performing and 8% (135) by women-performing Through the Kruskal-Wallis test (see Appendix A), we were able to un­ people. Thus, woman-performing people posted proportionately less derstand the statistical relevance of the relationship between the pictures-per-capita to the platform (i.e., they comprise 13% of the users occurrence of SM data coded with the different relational CES values and but posted 8% of the pictures). Among the 232 unique users (50% of the the respective distances to spatial predictors. Then, the Wilcoxon rank total) that shared residency information on their profile, 109 (47%) sum test (see Appendix B) allowed us to understand the relationship were coded as “local”, which amounted to 57,6% (514) of the total between relational CES values in their distance from each spatial pre­ pictures tagged with residency (892), while the remaining 53% as “non- dictor. Between the relational CES values that proved relevant in local”, contributing to 42,4% (378) of the pictures. explaining their distance from touristic amenities (see Fig. 3), the rela­ The chi-square test (see Appendix C) shows that the differences in tively longer distance that people walk on average to enjoy physical relational CES benefits uptake based on performed gender and residency recreation and existence value compared to enjoying cognitive value and are not random for some CES (see Figs. 5 and 6). built cultural heritage, resulted statistically relevant. Similarly, there is As expected, people performing as men appear to benefit at a statistical significance in the longer average distance walked from bus significantly higher rate than those performing as women from all the and train stations to enjoy physical recreation, natural cultural heritage and services which showed statistical correlation with this demographic social relations compared to that covered to benefit from experiential variable (see Appendix C and Fig. 5): physical recreation (p-value = recreation and built cultural heritage (see Fig. 3). Distance from walking 0.01392), experiential recreation (p-value = 0.03862), natural cultural paths is statistically explained by the enjoyment of spiritual value (p- heritage (p-value = 0.0001271), and social relations (p-value = 0.01641). value = 0.006275), built cultural heritage (p-value = 0.0007478) and In relation to place of residence, instead, relational CES benefits uptake physical recreation (p-value = 4.772e-05), being the respectively reveals a more equal distribution (see Fig. 6), with locals taking a higher decreasing average distance walked for their enjoyment also relevant number of pictures categorized as existence value (p-value = 0.004107), (see Fig. 3). Finally, the longer distance walked from hotspots to enjoy cognitive value (p-value = 0.0001788), natural cultural heritage (p-value spiritual value, cognitive value, social relations, and existence value = 2.8e-13), and – even if slightly – built cultural heritage (p-value <2.2e- compared to that covered to benefit from built cultural heritage showed to 16) compared to non-locals who, in turn, primarily appreciate landscape be statistically relevant as well (see Fig. 3). aesthetics (p-value <2.2e-16) and social relations (p-value = 3.077e-08). 3.2.2. Temporal predictors 3.2.4. Geo-temporal-demographic cross-analysis Temporal factors also seem to have an influence on the distribution After analysing how relational CES values distribution can be of the access to CES benefits. Most of the relational CES values which explained by the different predictors taken singularly, here we want to present a statistically significant correlation with the time of the day (see understand their combined effect. Statistically relevant correlations Appendix C), namely physical recreation (p-value = 2.84e-05), experien­ between seasons and performed gender (p-value = 0.0007301, see Ap­ tial recreation (p-value = 0.007496), existence value (p-value = 0.04913), pendix D), for example, show more nuanced relational CES benefit natural cultural heritage (p-value = 3.68e-10), landscape aesthetics (p- distribution, than the two predictors analysed separately (see Figs. 4 and value = 0.01897), and built cultural heritage (p-value = 3.147e-05) seem 5 respectively). to be enjoyed either in the morning or, mostly, in the afternoon (see In particular, summer and autumn seem to be the best or the worst Fig. 4), as confirmed by other studies (cfr. Bandara & Bandara, 2019). seasons to enjoy both physical recreation or social relations depending on Noticeably, and probably due to the relatively smaller sample of data whether you are a woman-performing or man-performing user respec­ coded as spiritual (p-value = 0.002116) and cognitive value (p-value = tively (with the only exception for physical recreation, which woman- 0.009655), a different timespan is preferred for their uptake, the former performing users enjoy at relatively high rate in autumn as well). being the most enjoyed during evenings and nights and the latter Similarly, woman-performing users show a preference for winter over showing comparable benefit uptake during mornings and afternoons. spring activities, as opposed to man-performing users mostly enjoying Moving up the temporal scale of the week, there is statistical sig­ springtime (see Fig. 7). nificance (see Appendix C) in people’s preference to enjoy experiential Given the statistical correlation between the weekly distribution of recreation (p-value = 0.0001113), cognitive value (p-value = 7.719e-05), relational CES values and residency (p-value = 2.441e-07), we notice natural cultural heritage (p-value = 1.138e-05), social relations (p-value = that both locals (more evidently) and non-locals seem to prefer to enjoy 0.006018) and built cultural heritage (p-value = 0.05525) over the time in the natural settings of Collserola during weekends, regardless of weekend. Interestingly, the slightly higher enjoyment of landscape aes­ the relational CES benefits enjoyed within those that showed statistical thetics on weekdays than on weekends is statistically relevant (p-value = correlation with both residency and day of the week (see Fig. 8 and 5.066e-07) (see Fig. 4). Appendix C). The only exception is given by non-locals, who seem to Looking at seasonal distribution, unexpectedly, CES depicting SM benefit more from landscape aesthetics and built cultural heritage during pictures are taken mainly in autumn and spring, fewer in winter and weekdays than weekends. even less in summer (see Fig. 4). This general trend is reflected in the The daily distribution of CES uptake is statistically correlated with annual distribution of the relational CES values statistically correlated both residency (p-value = 2.657e-09) and performed gender (p-value = 7 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Fig. 2. Distribution of relational Cultural Ecosystem Services (CES) values in Collserola park, Barcelona. Each pie chart shows the values distribution within the 124 grid cells in which the study area has been divided. The red graduated colour of each grid cell indicates the total amount of pictures geocodedwithin it. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 0.01027)(see Appendix D). Both locals and non-locals mostly enjoy with seasons (p-value = 1.248e-05) for the enjoyment of physical rec­ relational CES benefits in the afternoon. A clear increase in benefits reation and social relations (see Figs. 11 and Appendix A and C). While uptake from morning to afternoon is evident for all the relational CES performed gender does not seem to have an influence on the average that showed statistical correlation with both residency and time of the distance walked, man-performing users reveal to walk a wider range of day (see Fig. 9 and Appendix C) except for cognitive value, which locals distances from bus and train stations than woman-performing users to experience either in the morning or in the afternoon likewise, while non- enjoy every relevant relational CES benefits and during the whole day, locals seem to slightly prefer during mornings. Moreover, on average, with this difference being most visible at night. A similar result is shown non-locals seem to dedicate proportionally more evening- and night- by the distance walked to access relational CES benefits correlated with time than locals to the uptake of relational CES benefits. the seasonal distribution, in this case performed gender-based difference Physical recreation, experiential recreation and natural cultural heritage are mostly evident in autumn. resulted statistically correlated with both daytime and performed gender (see Fig. 10 and Appendix C). While the relational CES benefits 3.3. Reality check uptake revealed by woman-performing users decreases smoothly from mornings to evening and presents a steep decrease at night, man- Even though the number of respondents to the survey conducted in performing users’ enjoyment of relational CES benefits mostly hap­ Collserola (n = 1922) is significantly higher than the SM unique users in pens in the afternoon and at a lowering rate during mornings while this study, and the study design substantially different (no assessment of decreasing consistently during evening and night, but never reaching relational CES benefits uptake in the survey), the general trends and the drop experienced by people performing as women. comparable. The survey, in accordance with our results, reported the Finally, we analysed the interplay of all the different factors at once. highest frequency of visits in the south-western fringe of the park along Being both residency (p-value = 0.001856) and daytime (p-value = the famous walking path named Carretera de les Aigües, and mostly 2.346e-10) statistically correlated with distance from hotspots, we during spring and autumn, with a significant drop in summer. Weekends analysed the daily patterns of locals and non-locals’ uptake of the are also preferred by both survey respondents and SM users, but, ac­ relational CES values which resulted statistically correlated with the cording to the survey, people mostly visit the park in the morning (72%) combination of these geo-temporal-demographic variables, namely ex­ while SM shows a stronger preference in the afternoon. Among the istence and cognitive value, and built cultural heritage (see Fig. 11 and survey respondents, 75,2% would be coded as locals according to our Appendix A and C). In general, locals experience existence and cognitive classification, against the 47% found in our SM data sample, while value (apart from evenings for this latter) at shorter distances from regarding gender, the survey shows a similar but less stunning dispro­ hotspots than non-locals throughout the day. Finally, on average, locals portion (67% men vs 33% women) to that encountered in this study and non-locals move similar distances from hotspots to enjoy built cul­ (87% men-performing vs 13% women-performing). Also, when asked tural heritage, with non-locals reaching further distances than locals in about their habits in sharing their park experience on SM, they mostly absolute terms, especially during mornings and nights. mention tracking apps, such as Strava or Wikiloc, while Flickr does not Distance from bus and train stations correlates with performed appear to be a frequently used app. gender (p-value = 0.02073) and daytime (p-value = 0.001052) for the uptake of physical and experiential recreation, and natural cultural heritage, 8 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Fig. 3. a) Distance (m) between touristic amenities and geotagged SM data coded with statistically relevant explanatory relational CES values. b) Distance (m) between bus and train stations and geotagged SM data coded with statistically relevant explanatory relational CES values. c) Distance (m) between walking paths and geotagged SM data coded with statistically relevant explanatory relational CES values. 4) Distance (m) between hotspots and geotagged SM data coded with sta­ tistically relevant explanatory relational CES values. 4. Discussion above toward unearthing the relational dimensions of CES that are exposed and toward looking at the digital as both object and subject of This study aims to leverage the potential of SM data to generate inquiry (Elwood & Leszczynski, 2018). “snapshots of the city as a functioning system” (Crooks et al., 2010) and to gain novel insights into complex phenomena in social-ecological sys­ tems (Lopez et al., 2019) to inform justice and sustainability trans­ 4.1. Digital fine-scale traces of differentiated relational CES benefits formations. These aims speak to the CES community’s need to uptake understand the actual and fine-grained distribution of multiple and plural benefits and ascribed relational values in space and time, and to The analysis of social media metadata within the boundaries of the the feminist and digital geography community’s urge to investigate so­ peri-urban park of Collserola offers a privileged view of the relationship cial phenomena by blurring public/private and rean/more-than-real between multiple relational CES values and the differentiated distribu­ boundaries, politicizing life at the scale of the everyday and sensing tion across the several groups of urban dwellers interacting with and public embodied and situated perceptions through access to individuals’ within green spaces in the surroundings of densely populated cities. ideas and opinions (Koblet & Purves, 2020; Leszczynski, 2019; McLean, Spatial analyses confirm intuitive expectations that link the enjoy­ 2020; Morrow, Hawkins, & Kern, 2015). In the name of speaking to both ment of some relational CES benefits to specific spatial needs (e.g., of these communities, we target our discussion of the results presented walking relatively long distances from hotspots to enjoy social relations, spiritual, existence and cognitive value as well as moving further from 9 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Fig. 4. a) Percentage of daily variation in the statistically correlated uptake of relational CES benefits. b) Percentage of weekly variation in the statistically correlated uptake of relational CES benefits. c) Percentage of seasonal variation in the statistically correlated uptake of relational CES benefits. 10 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Fig. 5. Relative percentage of pictures distribution per performed gender and correlated relational CES values. Fig. 6. Relative percentage of pictures distribution per place of residence and correlated relational CES values. Fig. 7. Relative percentage of pictures seasonal distribution per performed gender and statistically correlated relational CES values. 11 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Fig. 8. Relative percentage of pictures weekly distribution per residency and statistically correlated relational CES values. touristic amenities and bus and train stations but closer to walking paths a new frontier for SM research, reflecting specific spatial and temporal for the uptake of physical recreation than for other benefits). The rela­ needs in relation to the social groups analysed (e.g., non-locals evading tively little distance travelled from walking paths and touristic amenities from crowded areas more than locals; reverse seasonal and daily pref­ (on average, <100 and 350 m respectively) confirms other studies erences in accessing relational CES benefits between woman- and man- findings (see Cao et al., 2022) and can be linked to several factors. First, performing users). given the mainly forestry cover, the enjoyable share of the park is of However, the different proportion in data availability per CES (e.g. relatively small size and, as expected, closely linked with paths and see Calcagni et al., 2022) and per predictor variable (as also warned in amenities. As well, the significantly dense network of walking paths that Shelton, 2022), may explain the lack of statistical significance for some eases access to the touristic amenities, especially Tibidabo mountain, of the relationships between the chosen variables and might have attracts both locals and tourists to the area. hampered the full potential of the proposed methodology, for instance Temporal analyses, in line with other studies, show predictable neglecting the peculiar spatial, temporal, and demographic patterns of trends of CES enjoyment during the day (e.g. most of the CES valued in some relational CES values. the morning and in the afternoon, fewer at evening and night) (cfr. Bandara & Bandara, 2019) and the week (e.g. overall, higher uptake over weekends than weekdays) (cfr., Havinga, Marcos, Bogaart, Hein, & 4.2. Embracing the fuzziness and challenging the boundaries of the more- Tuia, 2021), but rather surprising seasonal variability given the low than-real number of pictures taken in summer (differently from what found by da Silva Lopes, Cadima Remoaldo, & Ribeiro, 2018; but confirming find­ As feminist digital geography has recently asserted, local knowledge ings by Bandara & Bandara, 2019). This might be explained by the cannot be separated from the situated, embodied positionality of the relatively close beachline that could work as a higher attractor for both individual who produces or interprets it (Elwood, 2008; Leszczynski & tourists and local people in hot weather conditions. Elwood, 2015). Consequently – and in contrast to the objectivity, Finally, the fine-scale demographic characterization of SM data and neutrality and universality often ascribed to SM data by contested the multi-variate cross tabulation analyses performed in this study open algorithmic epistemology principles (Crawford, Miltner, & K., 2014; Leszczynski & Elwood, 2015; Shelton, 2022) – we acknowledge that the Fig. 9. Relative percentage of pictures daily distribution per residency and statistically correlated relational CES values. 12 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Fig. 10. Relative percentage of pictures daily distribution per performed gender and statistically related relational CES values. production of this data does not occur in a social vacuum, and that the reached conclusions different from this study, (e.g., Oteros-Rozas et al., digital environment is a place where subjectivities, everyday life and 2018 find SM to be mostly used by women), taking the context-specific perceptions of space are negotiated and forms of power reproduced dynamics revealed by SM data seriously, in a way that accounts for the (Armstrong et al., 2021; Elwood & Leszczynski, 2018; Stephens, 2013), data shortcomings without dismissing the imbalances exposed, is in a tangled process of online-offline behavioural and affective co- essential to furthering goals outlined in feminist cartography and data production of space and benefits uptake (Armstrong et al., 2021). In visualization agendas (Braidotti, 2005; D’Ignazio, 2015). this vein, scholars call for a feminist digital epistemology and method­ ology aimed at accounting for those processes, usually invisible in SM data assessment (Elwood & Leszczynski, 2018; Leszczynski & Elwood, 4.3. Opportunities and limitations 2015). To address the issue of SM data representativeness, mixed-method As seen, SM offer promising opportunities for studying the co- approaches have been proposed, using both SM and traditional data evolution of relational CES values and the ecosystem structures and (Ghermandi & Sinclair, 2019; Oteros-Rozas et al., 2018) or conducting functions to which they are ascribed (Blicharska et al., 2017). However, surveys to social media users – although with fairly low response rates the availability of SM data is dwindling and unstable. Platforms can shut (Lenormand et al., 2018). In our study, we relied on shared or revealed down for changing users demand, and access to data can be restricted information, which are also limited because contributed only by a due to data protection regulations (General Data Protection Regulation, portion of the SM users or assumed by the researcher. Procedure for 2016) or changed platforms’ Terms and Conditions of Use3, or denied to assessing representativeness may either reveal a general bias in SM data, avoid data manipulation in the corporate and political interests (Rich­ and thus confirm other evidence (Crampton et al., 2013; International terich, 2018). In addition, SM platforms access is unevenly distributed Telecommunication Union, 2016; Stephens, 2013), or unveil ground- across social groups. This can be due to continued issues of digital divide level spatial, temporal, and social structures and processes causing in­ regarding the access to devices, technologies, infrastructures and skills equalities in the case study area. In order to address aspects of injustices needed to make proper use of them (Ghermandi and Sinclair, 2019), or, in relational CES values and respective benefits access and uptake, those as suggested here, to the sense of safety felt by groups in cyberspace structures and processes need to be analysed as well (cf. Fortnam et al., (Adjuntament de Barcelona, 2018). 2019). Therefore, warranting ethical and inclusive practices in data use and For example, not accounting for the gender-imbalance found in both platform accessibility is essential to recover public trust and legitimacy SM and survey data might, on the one hand, underestimate the gendered in SM data research (de Juan et al., 2021; Ghermandi et al., 2023; geo-temporal patterns of access to relational CES benefits revealed by Hausmann et al., 2020). this study while, on the other, possibly neglect the effect of analogous In addition, we do not aim here to target or solve binary and heter­ differences in the predisposition of men and women (whether appearing onormative biases related with gender assumptions and inherent to or performing) to either respond to a survey or post on SM. This dif­ observatory methods (Fontán-Vela et al., 2021; McKenzie, Cohen, Seh­ ference can potentially be explained as fear for gender violence, which gal, Williamson, & Golinelli, 2006; Sebo, 2021) – and not yet sparing in Barcelona showed to mostly occur in public places for leisure and in active data collection methodologies either (see Valera & Casakin, 2022; the cyberspace (Adjuntament de Barcelona, 2018). Despite other studies Farías-Torbidoni and Morera Carbonell, 2020) – but this study is a first that used different sources of socio-demographic information obtained gender-neutral patterns of CES uptake (Martinez-Harms et al., 2018) or 3 For Flickr see https://www.flickr.com/help/terms 13 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Fig. 11. a) Daily distribution of residency-based distances from hotspots for the uptake of the different statistically correlated relational CES benefits. b) Daily distribution of performed gender-based distance from bus and train stations for the uptake of the different statistically correlated relational CES benefits. c) Seasonal distribution of performed gender-based distance from bus and train stations for the uptake of the different statistically correlated relational CES benefits. 14 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 attempt to engage with data skewedness in terms of gender and place of Flickr’s terms of service. To account for users’ privacy, in the retrieval residence, as called for in recent literature (Ghermandi et al., 2023). As process only performed gender (woman/man) and translated places of future research prospects, we encourage stated gender data collections residence (local/non-local) were noted down for each ID. The most to further analyse how gender identity is performed online and weights sensitive of these data, such as names and places of residence, were not on individual behaviours, both in digital and public green spaces. recorded. To reinforce the anonymity guaranty to Flickr users, we Finally, we did not perform intersectional analysis (namely combining encrypted the folders storing the data. Individual data and spatial data the two demographic variables) because not statistically correlated, but have been represented in aggregated form and disjunct from each other, we encourage other studies to engage in investigating all the possible to avoid the possible reconstruction of links between them through data statistically relevant combinations of the different dimensions, also processing. It should be noted that the retrieval and use of SM data re­ including other demographic factors, such as age, class, race, disability, quires strong ethical considerations (Boyd & Crawford, 2012; Gher­ education level, etc. mandi et al., 2023; Ghermandi & Sinclair, 2019). Especially with reference to personal data, retrieved either manually or using advanced 5. Conclusions methods such as ‘inferring demographics’ (see, i.e., Lampos, Aletras, Geyti, Zou, & Cox, 2016), the anonymity of users needs to be guaranteed With the aim to inform value-based justice and sustainability trans­ by applying privacy precautions in observation of the respective legis­ formations, this study provides several insights by complementing SM lative body (General Data Protection Regulation, 2016). Concerns on data content analysis with metadata analysis. Firstly, it shows the dif­ whether SM data should be considered useful to research in the public ferential access to CES benefits across space, time, and demographics. In interest underlies an ongoing ethical debate that, to date, restricts data particular, we show that the qualitative information provided by SM access and hinders data sharing by users. The scientific relevance that metadata is crucial to advancing the understanding of relational CES information extracted from SM increasingly demonstrates, call for ur­ values that tie people to natural environments, e.g., by analysing the gently establishing a common and clear ethical framework. relative distribution of relational CES values held by different social actors across space and time. Overall, the revealed patterns correspond to those resulting from an extensive survey conducted in the same case- Declaration of Competing Interest study area, confirming that social media platforms are useful and valid data sources to assess relational CES values distribution at fine-grained The authors declare that they have no known competing financial and wide scales. interests or personal relationships that could have appeared to influence Secondly, with the intention to account for the embodied position­ the work reported in this paper. ality of SM users in relational CES values distribution over space and time, we performed a descriptive cross-analysis and demonstrated how Acknowledgements embodied experiences of CES benefits are dynamically negotiated across the demographic categories that was possible to retrieve or assume. We, We thank Margarita Triguero-Mas, Amalia Calderón-Argelich and thus, provide novel information about the multiple and situated spatial Andréanne Breton-Carbonneau for their valuable contribution to the and temporal ordinary practices and tactics of CES benefits uptake, and improvement of the manuscript. FC thanks the AGAUR Catalan the respective relational values distribution, producing on-the-ground governmental agency (Grant number 2018FI_B00635) and the German- evidence for informing transformations rooted in CES and ES provi­ Israeli Foundation (GIF grant no. I-1533-500.15/2021) for the funding sion that are sensitive to diversity (Beebeejaun, 2017; Connolly, 2019). received to support this study. JL acknowledges additional funding from At the same time, this study proves the importance of considering SM the European Research Council (ERC Consolidator Grant: 818002- platforms as objects of study as well, revealing neglected relationships URBAG). J.J.T.C.’s contribution was supported by the Spanish Ministry between digital and offline reality. In particular, it highlights an of Sciences, Innovation, and University’s Subprogram of Juan de la important aspect to always consider when performing social media- Cierva Incoporacion (IJCI-2016-31100) and through the European based CES assessments: the representativeness of the SM sample for Research Council (Greenlulus 678034). This research also contributes to the specific case-study area and its different social groups. By revealing the Maria de Maetzu Unit of Excellence grant (MDM-2015-0552) at demographic information of SM users, we addressed a crucial gap in ICTA-UAB. most SM-based CES studies and demonstrated the uneven proportion of performing men and women in this study, partially confirmed by the Supplementary data survey as well. Although we cannot assume it is fully representative of the offline reality, geolocated SM data (in this case Flickr) reveal im­ Supplementary data to this article can be found online at https://doi. balances in gender representation, and representation at the intersection org/10.1016/j.diggeo.2023.100065. with other demographic factors, that are critical for justice trans­ formations to account for. However, SM-based assessment approaches References might lead to highly inaccurate representations of the actual access to and uptake of relational CES values in place if the limitations of the data Adjuntament de Barcelona. (2018). Enquesta de Violència Masclista a Catalunya. Edició highlighted here are not accounted for, and would thus not serve to 2016. Alieva, D., Holgado, D., de Juan, S., Ruiz-Frau, A., Villasante, S., & Maya-Jariego, I. properly guide decision-making. In addition, in line with other studies (2021). Assessing landscape features and ecosystem services of marine protected (Ghermandi & Sinclair, 2019; Vaz et al., 2020), the outcomes of this areas through photographs on social media: Comparison of two archipelagos in study seriously question the wider application of SM-based assessments Spain. Environment, Development and Sustainability, 24(7), 9623–9641. https://doi. org/10.1007/s10668-021-01841-y without complementing and validating them with ground-level data. Ames, M., & Naaman, M. (2007). Why we tag: Motivations for annotation in mobile and Notwithstanding, statistically relevant performing gender- and online media. In , 971–980. Proceedings of the SIGCHI Conference on Human Factors in residency-based differences in the accessibility to Collserola, evident Computing Systems. https://doi.org/10.1145/1240624.1240772 from our results, also indicate that demographic-specific approaches to Amorim-Maia, A. T., Calcagni, F., John, J., Connolly, T., Anguelovski, I., & Langemeyer, J. (2020). Hidden drivers of social injustice: Uncovering unequal urban and peri-urban landscape planning are overdue. cultural ecosystem services behind green gentrification. Environmental Science and Policy, 112(May), 254–263. https://doi.org/10.1016/j.envsci.2020.05.021 Ethics statement Andersson, E., Langemeyer, J., Borgström, S., McPhearson, T., Haase, D., Kronenberg, J., … Baró, F. (2019). Enabling green and blue infrastructure to improve contributions to human well-being and equity in Urban Systems. BioScience, 69(7), 566–574. Flickr data were collected through the Flickr API complying with https://doi.org/10.1093/biosci/biz058 15 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Andersson, E., Tengö, M., McPhearson, T., & Kremer, P. (2015). Cultural ecosystem Comissió institucional del Pla especial de Collserola. (2019). pla Especial de protecció del services as a gateway for improving urban sustainability. Ecosystem Services, 12, medi natural i del paisatge del parc natural de la serra de collserola. 165–168. https://doi.org/10.1016/j.ecoser.2014.08.002 Connolly, J. J. T. (2019). From Jacobs to the just City: A foundation for challenging the Angradi, T. R., Launspach, J. J., & Debbout, R. (2018). Determining preferences for green planning orthodoxy. Cities, 91(September 2017), 64–70. https://doi.org/ ecosystem benefits in Great Lakes areas of concern from photographs posted to social 10.1016/j.cities.2018.05.011 media. Journal of Great Lakes Research, 44(2), 340–351. https://doi.org/10.1016/j. Crampton, J. W., Graham, M., Poorthuis, A., Shelton, T., Stephens, M., Wilson, M. W., & jglr.2017.12.007 Zook, M. (2013). Beyond the geotag: Situating ‘big data’ and leveraging the potential Arias-Arévalo, P., Gómez-Baggethun, E., Martín-López, B., & Pérez-Rincón, M. (2018). of the geoweb. Cartography and Geographic Information Science, 40(2), 130–139. Widening the evaluative space for ecosystem services: A taxonomy of plural values https://doi.org/10.1080/15230406.2013.777137 and valuation methods. Environmental Values, 27(1), 29–53. https://doi.org/ Crawford, K. G., Miltner, M. L., & K.. (2014). Big data| critiquing big data: Politics, ethics, 10.3197/096327118X15144698637513 epistemology| special section introduction. International Journal of Communication, 8, Armstrong, M., Derrien, M. M., & Schaefer-Tibbett, H. (2021). The dynamics of trail use 1663–1672. and trip reporting: Understanding visitor experiences within social-ecological Crooks, A. T., Croitoru, A., Jenkins, A., Mahabir, R., Agouris, P., & Stefanidis, A. (2010). systems. Journal of Outdoor Recreation and Tourism, 100456. https://doi.org/ User-generated big data and urban morphology. Built Environment, 42(3). 10.1016/j.jort.2021.100456 Currie, M. J. B., Lackova, P., & Dinnie, E. (2016). Greenspace matters: Exploring links Arts, I., Fischer, A., Duckett, D., & Wal, R. (2021). The Instagrammable outdoors – between greenspace, gender and well-being with conservation volunteers. Landscape Investigating the sharing of nature experiences through visual social media. People Research, 41(6), 641–651. https://doi.org/10.1080/01426397.2016.1208813 and Nature, 3(6), 1244–1256. https://doi.org/10.1002/pan3.10239 D’Ignazio, C. (2015). What would feminist data visualization look like? MIT Center for Bandara, T., & Bandara, T. P. (2019). Whale watching in Sri Lanka: Understanding the Civic Media, 1988, 1–9. metadata of crowd-sourced photographs on FlickrTM social media platform. Sri Danford, R. S., Strohbach, M. W., Ryan, R., Nicolson, C., & Warren, P. S. (2014). What Lanka Journal of Aquatic Sciences, 24(2), 41. https://doi.org/10.4038/sljas. does it take to achieve equitable urban tree canopy distribution? A Boston case v24i2.7566 study. Cities and the Environment, 7(1). Article 2. Baró, P. F., Langemeyer, J., Łaszkiewicz, E., & Kabisch, N. (2021). Editorial to the special Derungs, C., & Purves, R. S. (2016). Characterising landscape variation through spatial issue“advancing urban ecosystem service implementation and assessment folksonomies. Applied Geography, 75, 60–70. https://doi.org/10.1016/j. considering different dimensions of environmental justice. Environmental Science & apgeog.2016.08.005 Policy, 115, 43–46. https://doi.org/10.1016/j.envsci.2020.10.008 Díaz, S., Demissew, S., Carabias, J., Joly, C., Lonsdale, M., Ash, N., … Zlatanova, D. Barros, C., Moya-Gómez, B., & García-Palomares, J. C. (2019). Identifying temporal (2015). The IPBES conceptual framework—Connecting nature and people. Current patterns of visitors to National Parks through geotagged photographs. Sustainability, Opinion in Environmental Sustainability, 14, 1–16. https://doi.org/10.1016/j. 11(24), 6983. https://doi.org/10.3390/su11246983 cosust.2014.11.002 Barry, S. J. (2014). Using social media to discover public values, interests, and Dickinson, D. C., & Hobbs, R. J. (2017). Cultural ecosystem services: Characteristics, perceptions about cattle grazing on park lands. Environmental Management, 53(2), challenges and lessons for urban green space research. Ecosystem Services, 25, 454–464. https://doi.org/10.1007/s00267-013-0216-4 179–194. https://doi.org/10.1016/j.ecoser.2017.04.014 Beebeejaun, Y. (2017). Gender, urban space, and the right to everyday life. Journal of Ding, K., Yang, M., & Luo, S. (2021). Mountain landscape preferences of millennials Urban Affairs, 39(3), 323–334. https://doi.org/10.1080/07352166.2016.1255526 based on social media data: A case study on Western Sichuan. Land, 10(11), 1246. Bernetti, I., Chirici, G., & Sacchelli, S. (2019). Big data and evaluation of cultural https://doi.org/10.3390/land10111246 ecosystem services: An analysis based on geotagged photographs from social media van Doorn, N. (2011). Digital spaces, material traces: How matter comes to matter in in Tuscan forest (Italy). IForest - Biogeosciences and Forestry, 12(1), 98–105. https:// online performances of gender, sexuality and embodiment. Media, Culture and doi.org/10.3832/ifor2821-011 Society, 33(4), 531–547. https://doi.org/10.1177/0163443711398692 Blicharska, M., Smithers, R. J., Hedblom, M., Hedenås, H., Mikusiński, G., Pedersen, E., Elwood, S. (2008). Volunteered geographic information: Future research directions … Svensson, J. (2017). Shades of grey challenge practical application of the cultural motivated by critical, participatory, and feminist GIS. GeoJournal, 72(3–4), 173–183. ecosystem services concept. Ecosystem Services, 23(November 2016), 55–70. https:// https://doi.org/10.1007/s10708-008-9186-0 doi.org/10.1016/j.ecoser.2016.11.014 Elwood, S. (2021). Digital geographies, feminist relationality, black and queer code Boyd, D., & Crawford, K. (2012). Critical questions for big data: Provocations for a studies: Thriving otherwise. Progress in Human Geography, 45(2), 209–228. https:// cultural, technological, and scholarly phenomenon. Information, Communication & doi.org/10.1177/0309132519899733 Society, 15(5), 662–679. https://doi.org/10.1080/1369118X.2012.678878 Elwood, S., & Leszczynski, A. (2018). Feminist digital geographies. Gender, Place and Braidotti, R. (2005). A critical cartography of feminist post-postmodernism. Australian Culture, 25(5), 629–644. https://doi.org/10.1080/0966369X.2018.1465396 Feminist Studies, 20(47), 169–180. https://doi.org/10.1080/08164640500090319 Ernstson, H. (2013). The social production of ecosystem services: A framework for Büscher, B., Koot, S., & Nelson, I. L. (2017). Introduction. Nature 2.0: New media, online studying environmental justice and ecological complexity in urbanized landscapes. activism and the cyberpolitics of environmental conservation. Geoforum, 79, Landscape and Urban Planning, 109(1), 7–17. https://doi.org/10.1016/j. 111–113. https://doi.org/10.1016/j.geoforum.2016.12.001 landurbplan.2012.10.005 Byron, P. (2020). Digital media, friendship and cultures of care. Routledge. European Commission. (2020). EU biodiversity strategy for 2030—Bringing nature back into Calcagni, F., Amorim Maia, A. T., Connolly, J. J. T., & Langemeyer, J. (2019). Digital co- our lives. construction of relational values: Understanding the role of social media for Farías-Torbidoni, E. I., & Morera Carbonell, S. (2020). Estudi d’afluència, freqüentació i sustainability. Sustainability Science, 1–13. https://doi.org/10.1007/s11625-019- caracterització dels usuaris- visitants del parc natural de la serra de collserola (Issue July). 00672-1 https://doi.org/10.13140/RG.2.2.27890.12485 Calcagni, F., Nogué Batallé, J., Baró, F., & Langemeyer, J. (2022). A tag is worth a Fischer, A., & Eastwood, A. (2016). Coproduction of ecosystem services as human-nature thousand pictures: A framework for an empirically grounded typology of relational interactions-an analytical framework. Land Use Policy, 52, 41–50. https://doi.org/ values through social media. Ecosystem Services, 58, Article 101495. https://doi.org/ 10.1016/j.landusepol.2015.12.004 10.1016/j.ecoser.2022.101495 Fish, R., Church, A., Willis, C., Winter, M., Tratalos, J. A., Haines-Young, R., & Calderón-Argelich, A., Benetti, S., Anguelovski, I., Connolly, J. J. T., Langemeyer, J., & Potschin, M. (2016). Making space for cultural ecosystem services: Insights from a Baró, F. (2021). Tracing and building up environmental justice considerations in the study of the UK nature improvement initiative. Ecosystem Services, 21(August), urban ecosystem service literature: A systematic review. Landscape and Urban 329–343. https://doi.org/10.1016/j.ecoser.2016.09.017 Planning, 214(October). https://doi.org/10.1016/j.landurbplan.2021.104130 Fish, R., Church, A., & Winter, M. (2016). Conceptualising cultural ecosystem services: A Cao, H., Wang, M., Su, S., & Kang, M. (2022). Explicit quantification of coastal cultural novel framework for research and critical engagement. Ecosystem Services, 21 ecosystem services: A novel approach based on the content and sentimental analysis (January 2015), 208–217. https://doi.org/10.1016/j.ecoser.2016.09.002 of social media. Ecological Indicators, 137, Article 108756. https://doi.org/10.1016/ Fontán-Vela, M., Rivera-Navarro, J., Gullón, P., Díez, J., Anguelovski, I., & Franco, M. j.ecolind.2022.108756 (2021). Active use and perceptions of parks as urban assets for physical activity: A Chan, K. M. A., Balvanera, P., Benessaiah, K., Chapman, M., Díaz, S., Gómez- mixed-methods study. Health & Place, 71, Article 102660. https://doi.org/10.1016/ Baggethun, E., … Turner, N. (2016). Why protect nature? Rethinking values and the j.healthplace.2021.102660 environment. Proceedings of the National Academy of Sciences, 113(6), 1462–1465. Fortnam, M., Brown, K., Chaigneau, T., Crona, B., Daw, T. M., Gonçalves, D., … Schulte- https://doi.org/10.1073/pnas.1525002113 Herbruggen, B. (2019). The gendered nature of ecosystem services. Ecological Chan, K. M. A., Gould, R. K., & Pascual, U. (2018). Editorial overview: Relational values: Economics, 159(February), 312–325. https://doi.org/10.1016/j. What are they, and what’s the fuss about? Current Opinion in Environmental ecolecon.2018.12.018 Sustainability, 35(December), A1–A7. https://doi.org/10.1016/j.cosust.2018.11.003 Geneletti, D. (Ed.). (2016). Handbook on biodiversity and ecosystem services in impact Chan, K. M. A., Guerry, A. D., Balvanera, P., Klain, S., Satterfield, T., Basurto, X., … assessment. Hannahs, N. (2012). Where are cultural and social in ecosystem services? A General Data Protection Regulation. (2016). Testimony of European Parliament. framework for constructive engagement. BioScience, 62(8), 744–756. https://doi. Ghermandi, A., Camacho-valdez, V., & Trejo-espinosa, H. (2020). Social media-based org/10.1525/bio.2012.62.8.7 analysis of cultural ecosystem services and heritage tourism in a coastal region of Chan, K. M. A., Satterfield, T., & Goldstein, J. (2012). Rethinking ecosystem services to Mexico. Tourism Management, 77(September 2019), Article 104002. https://doi.org/ better address and navigate cultural values. Ecological Economics, 74, 8–18. https:// 10.1016/j.tourman.2019.104002 doi.org/10.1016/j.ecolecon.2011.11.011 Ghermandi, A., Langemeyer, J., van Berkel, D. B., Calcagni, F., Depietri, Y., Egarter Clemente, P., Calvache, M., Antunes, P., Santos, R., Cerdeira, J. O., & Martins, M. J. Vigl, L., … Wood, S. A. (2023). Social media data for environmental sustainability: A (2019). Combining social media photographs and species distribution models to map critical review of opportunities, threats and ethical use. One, Earth, 6(3), 236–250. cultural ecosystem services: The case of a Natural Park in Portugal. Ecological https://doi.org/10.1016/j.oneear.2023.02.008 Indicators, 96, 59–68. https://doi.org/10.1016/j.ecolind.2018.08.043 16 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Ghermandi, A., & Sinclair, M. (2019). Passive crowdsourcing of social media in Lazer, D., Hargittai, E., Freelon, D., Gonzalez-bailon, S., & Munger, K. (2021). environmental research: A systematic map. Global Environmental Change, 55, 36–47. Meaningful measures of human society in the twenty-first century. Nature, 595 https://doi.org/10.1016/j.gloenvcha.2019.02.003 (July), 189–196. https://doi.org/10.1038/s41586-021-03660-7 Gliozzo, G., Pettorelli, N., & Haklay, M. (2016). Using crowdsourced imagery to detect Lenormand, M., Luque, S., Langemeyer, J., Tenerelli, P., Zulian, G., Aalders, I., … cultural ecosystem services: A case study in South Wales, UK. Ecology and Society, 21 Woods, H. (2018). Multiscale socio-ecological networks in the age of information. (3). https://doi.org/10.5751/es-08436-210306 PLoS One, 13(11), 1–16. https://doi.org/10.1371/journal.pone.0206672 Gosal, A. S., Geijzendorffer, I. R., Václavík, T., Poulin, B., & Ziv, G. (2019). Using social Leszczynski, A. (2019). Digital methods III: The digital mundane. Progress in Human media, machine learning and natural language processing to map multiple Geography, 44(6), 1194–1201. https://doi.org/10.1177/0309132519888687 recreational beneficiaries. Ecosystem Services, 38(June 2018), Article 100958. Leszczynski, A., & Elwood, S. (2015). Feminist geographies of new spatial media. https://doi.org/10.1016/j.ecoser.2019.100958 Canadian Geographer, 59(1), 12–28. https://doi.org/10.1111/cag.12093 Gosal, A. S., & Ziv, G. (2020). Landscape aesthetics: Spatial modelling and mapping using Lopez, B. E., Magliocca, N. R., & Crooks, A. T. (2019). Challenges and opportunities of social media images and machine learning. Ecological Indicators, 117(May), 106638. social media data for socio-environmental systems research. Land, 8(7), 1–18. https://doi.org/10.1016/j.ecolind.2020.106638 https://doi.org/10.3390/land8070107 Grossmann, K., Connolly, J. J. T., Mattioli, G., & Nitschke, L. (2022). From sustainable Luo, F., Cao, G., Mulligan, K., & Li, X. (2016). Explore spatiotemporal and demographic development to social- ecological justice: Addressing taboos and naturalizations in characteristics of human mobility via twitter: A case study of Chicago. Applied order to shift perspective. EPE: Nature and Space, 0(0), 1–23. https://doi.org/ Geography, 70, 11–25. https://doi.org/10.1016/j.apgeog.2016.03.001 10.1177/25148486211029427 Mancini, F., Coghill, G. M., & Lusseau, D. (2018). Using social media to quantify spatial Guo, Z., Zhang, L., & Li, Y. (2010). Increased dependence of humans on ecosystem and temporal dynamics of nature-based recreational activities. PLoS One, 13(7), services and biodiversity. PLoS One, 5(10). https://doi.org/10.1371/journal. 1–19. https://doi.org/10.1371/journal.pone.0200565 pone.0013113 Martinez-Alier, J., Kallis, G., Veuthey, S., Walter, M., & Temper, L. (2010). Social Haines-Young, R., & Potschin, M. (2018). Common international classification of ecosystem metabolism, ecological distribution conflicts, and valuation languages. Ecological services (CICES) V5.1 guidance on the application of the revised structure. Economics, 70(2), 153–158. https://doi.org/10.1016/j.ecolecon.2010.09.024 Hamstead, Z. A., Fisher, D., Ilieva, R. T., Wood, S. A., McPhearson, T., & Kremer, P. Martinez-Harms, M. J., Bryan, B. A., Wood, S. A., Fisher, D. M., Law, E., Rhodes, J. R., … (2018). Geolocated social media as a rapid indicator of park visitation and equitable Wilson, K. A. (2018). Inequality in access to cultural ecosystem services from park access. Computers, Environment and Urban Systems, 72(January), 38–50. https:// protected areas in the Chilean biodiversity hotspot. Science of the Total Environment, doi.org/10.1016/j.compenvurbsys.2018.01.007 636, 1128–1138. https://doi.org/10.1016/j.scitotenv.2018.04.353 Haraway, D. (1988). Situated knowledges: The science question in feminism and the Massey, D. (1994). Space, place, and gender. Polity press. privilege of partial perspective. Feminist Studies, 14(3), 575. https://doi.org/ McKenzie, T. L., Cohen, D. A., Sehgal, A., Williamson, S., & Golinelli, D. (2006). System 10.2307/3178066 for observing play and recreation in communities (SOPARC): Reliability and Harvey, D. (1996). Justice, nature & the geography of difference. Blackwell Publishing. feasibility measures. Journal of Physical Activity and Health, 3(s1), S208–S222. Hausmann, A., Toivonen, T., Fink, C., Heikinheimo, V., Kulkarni, R., Tenkanen, H., & Di https://doi.org/10.1123/jpah.3.s1.s208 Minin, E. (2020). Understanding sentiment of national park visitors from social McLean, J. (2020). Changing digital geographies—Technologies, environments and people. media data. People and Nature, 2(3), 750–760. https://doi.org/10.1002/pan3.10130 Palgrave macmillan. Havinga, I., Bogaart, P. W., Hein, L., & Tuia, D. (2020). Defining and spatially modelling MEA. (2005). Ecosystems and human well-being—Synthesis. In , 1. The Millennium cultural ecosystem services using crowdsourced data. Ecosystem Services, 43. https:// Ecosystem Assessment series. https://doi.org/10.1007/BF02987493 doi.org/10.1016/j.ecoser.2020.101091 Meikle, G. (2016). Social media—Communication, sharing and visibility (p. 147). Havinga, I., Marcos, D., Bogaart, P. W., Hein, L., & Tuia, D. (2021). Social media and Routledge, Taylor & Francis Group. deep learning capture the aesthetic quality of the landscape. Scientific Reports, 11(1), Milcu, A. I. H., Hanspach, J., Abson, D., & Fischer, J. (2013). Cultural ecosystem services: 20000. https://doi.org/10.1038/s41598-021-99282-0 A literature review and prospects for future research. Ecology and Society, 18(3), 44. Himes, A., & Muraca, B. (2018). Relational values: The key to pluralistic valuation of Morrow, O., Hawkins, R., & Kern, L. (2015). Feminist research in online spaces. Gender, ecosystem services. Current Opinion in Environmental Sustainability, 35(March 2019), Place and Culture, 22(4), 526–543. https://doi.org/10.1080/ 1–7. https://doi.org/10.1016/j.cosust.2018.09.005 0966369X.2013.879108 Huai, S., Chen, F., Liu, S., Canters, F., & Van de Voorde, T. (2022). Using social media Muraca, B. (2016). Re-appropriating the ecosystem services concept for a decolonization photos and computer vision to assess cultural ecosystem services and landscape of ‘nature’. In Nature and experience—Phenomenology and the environment (pp. features in urban parks. Ecosystem Services, 57, Article 101475. https://doi.org/ 143–156). Bannon, B. 10.1016/j.ecoser.2022.101475 Nelson, I. L., Hawkins, R., & Govia, L. (2022). Feminist digital natures. Environment and Huang, H., Gartner, G., & Turdean, T. (2013). Social media data as a source for studying Planning E: Nature and Space. https://doi.org/10.1177/25148486221123136, people’s perception and knowledge of environments. Mitteilungen Der 251484862211231. Österreichischen Geographischen Gesellschaft, 155, 291–302. Oteros-Rozas, E., Martín-López, B., Fagerholm, N., Bieling, C., & Plieninger, T. (2018). Ilieva, R. T., & McPhearson, T. (2018). Social-media data for urban sustainability. Nature Using social media photos to explore the relation between cultural ecosystem Sustainability, 1(10), 553–565. https://doi.org/10.1038/s41893-018-0153-6 services and landscape features across five European sites. Ecological Indicators, 94 International Telecommunication Union. (2016). Measuring the information society (2), 74–86. https://doi.org/10.1016/j.ecolind.2017.02.009 report 2016. In International Telecommunication Union. https://doi.org/10.3359/ Panelli, R. (2010). More-than-human social geographies: Posthuman and other oz0303157 possibilities. Progress in Human Geography, 34(1), 79–87. https://doi.org/10.1177/ IPBES. (2018). Summary for policymakers of the regional assessment report on biodiversity 0309132509105007 and ecosystem Services for Europe and Central Asia (issue July). https://doi.org/ Pascual, U., Balvanera, P., Díaz, S., Pataki, G., Roth, E., Stenseke, M., … Yagi, N. (2017). 10.13140/RG.2.2.32932.58248 Valuing nature’s contributions to people: The IPBES approach. Current Opinion in Jacobs, S., Martín-López, B., Barton, D. N., Dunford, R., Harrison, P. A., Kelemen, E., … Environmental Sustainability, 26–27, 7–16. https://doi.org/10.1016/j. Smith, R. (2018). The means determine the end – Pursuing integrated valuation in cosust.2016.12.006 practice. Ecosystem Services, 29, 515–528. https://doi.org/10.1016/j. Pastur, G. M., Peri, P. L., Lencinas, M. V., Garcı-Llorente, M., & Martìn-Lopez, B. (2016). ecoser.2017.07.011 Spatial patterns of cultural ecosystem services provision in Southern Patagonia. https:// de Juan, S., Ospina-Álvarez, A., Villasante, S., & Ruiz-Frau, A. (2021). A graph theory doi.org/10.1007/s10980-015-0254-9 approach to assess nature’s contribution to people at a global scale. Scientific reports, Pilcher, J. (2017). Names and “doing gender”: How forenames and surnames contribute 256991401 bytes. https://doi.org/10.1038/s41598-021-88745-z to gender identities, difference, and inequalities. Sex Roles, 77(11− 12), 812–822. Koblet, O., & Purves, R. S. (2020). From online texts to landscape character assessment: https://doi.org/10.1007/s11199-017-0805-4 Collecting and analysing first-person landscape perception computationally. Plieninger, T., Bieling, C., Fagerholm, N., Byg, A., Hartel, T., Hurley, P., … Huntsinger, L. Landscape and Urban Planning, 197, Article 103757. https://doi.org/10.1016/j. (2015). The role of cultural ecosystem services in landscape management and landurbplan.2020.103757 planning. Current Opinion in Environmental Sustainability, 14, 28–33. https://doi.org/ Krange, O., & Skogen, K. (2007). Reflexive tradition: Young working–class hunters 10.1016/j.cosust.2015.02.006 between wolves and modernity. YOUNG, 15(3), 215–233. https://doi.org/10.1177/ Plieninger, T., Dijks, S., Oteros-Rozas, E., & Bieling, C. (2013). Assessing, mapping and 110330880701500301 quantifying cultural ecosystem services at community level. Land Use Policy, Kremer, P., Hamstead, Z., Haase, D., McPhearson, T., Frantzeskaki, N., Andersson, E., … 118–129. Elmqvist, T. (2016). Key insights for the future of urban ecosystem services research. Quercia, D., & Saez, D. (2014). Mining urban deprivation from foursquare: Implicit Ecology and Society, 21(2). https://doi.org/10.5751/ES-08445-210229 crowdsourcing of city land use. IEEE Pervasive Computing, 13(2), 30–36. https://doi. Lampos, V., Aletras, N., Geyti, J. K., Zou, B., & Cox, I. J. (2016). Inferring the org/10.1109/MPRV.2014.31 Socioeconomic Status of Social Media Users Based on Behaviour and Language. In Richards, D. R., & Friess, D. A. (2015). A rapid indicator of cultural ecosystem service N. Ferro, F. Crestani, M.-F. Moens, J. Mothe, F. Silvestri, G. M. Di Nunzio, … usage at a fine spatial scale: Content analysis of social media photographs. Ecological G. Silvello (Eds.), 9626. Advances in Information Retrieval: 38th European Conference Indicators, 53, 187–195. https://doi.org/10.1016/j.ecolind.2015.01.034 on IR Research, ECIR 2016, Padua, Italy, March 20–23, 2016. Proceedings (pp. Richterich, A. (2018). How data-driven research fuelled the Cambridge Analitica controversy 689–695). Springer, Lecture Notes in Computer Science. https://doi.org/10.1007/ (p. 7623). https://doi.org/10.1285/i20356609v11i2p528 978-3-319-30671-1. Riechers, M., Barkmann, J., & Tscharntke, T. (2016). Perceptions of cultural ecosystem Langemeyer, J., & Connolly, J. J. T. (2020). Weaving notions of justice into urban services from urban green. Ecosystem Services, 17, 33–39. https://doi.org/10.1016/j. ecosystem services research and practice. Environmental Science and Policy, 109 ecoser.2015.11.007 (September 2019), 1–14. https://doi.org/10.1016/j.envsci.2020.03.021 Rogg Korsvik, T., & Rustad, L. M. (2015). What is the gender dimension in research? Case studies in interdisciplinary research. Kilden. 17 F. Calcagni et al. Digital Geography and Society 5 (2023) 100065 Ruiz-Frau, A., Ospina-Alvarez, A., Villasante, S., Pita, P., Maya-Jariego, I., & de Juan, S. Stephens, M. (2013). Gender and the GeoWeb: Divisions in the production of user- (2020). Using graph theory and social media data to assess cultural ecosystem generated cartographic information. GeoJournal, 78(6), 981–996. https://doi.org/ services in coastal areas: Method development and application. Ecosystem Services, 10.1007/s10708-013-9492-z 45(July), Article 101176. https://doi.org/10.1016/j.ecoser.2020.101176 Tadaki, M., Sinner, J., & Chan, K. M. A. (2017). Making sense of environmental values: A Schulz, C., & Martin-Ortega, J. (2018). Quantifying relational values—Why not? Current typology of concepts. Ecology and Society, 22(1). https://doi.org/10.5751/ES-08999- Opinion in Environmental Sustainability, 35(December 2017), 15–21. https://doi.org/ 220107 10.1016/j.cosust.2018.10.015 Tenkanen, H., Di Minin, E., Heikinheimo, V., Hausmann, A., Herbst, M., Kajala, L., & Sebo, P. (2021). Performance of gender detection tools: A comparative study of name-to- Toivonen, T. (2017). Instagram, Flickr, or twitter: Assessing the usability of social gender inference services. Journal of the Medical Library Association, 109(3). https:// media data for visitor monitoring in protected areas. Scientific Reports, 7(1), 1–11. doi.org/10.5195/jmla.2021.1185 https://doi.org/10.1038/s41598-017-18007-4 Shelton, T. (2022). Situated mapping: Visualizing urban inequality between the god trick Tieskens, K. F., Van Zanten, B. T., Schulp, C. J. E., & Verburg, P. H. (2018). Aesthetic and strategic positivism. ACME: An International Journal for Critical Geographies, 21 appreciation of the cultural landscape through social media: An analysis of revealed (4), 346–356. https://doi.org/10.31235/osf.io/8zswy preference in the Dutch river landscape. Landscape and Urban Planning, 177(May), Shelton, T. (2023). Challenging opacity, embracing fuzziness: Geographical thought and 128–137. https://doi.org/10.1016/j.landurbplan.2018.05.002 praxis in a post-truth age. Dialogues in Human Geography. https://doi.org/10.1177/ Valera, S., & Casakin, H. (2022). Integrating observation and network analysis to identify 20438206231157891, 204382062311578. patterns of use in the public space: A gender perspective. Frontiers in Psychology, 13, Shelton, T., Poorthuis, A., & Zook, M. (2015). Social media and the city: Rethinking Article 898809. https://doi.org/10.3389/fpsyg.2022.898809 urban socio-spatial inequality using user-generated geographic information. Vaz, A. S., Moreno-Llorca, R. A., Gonçalves, J. F., Vicente, J. R., Méndez, P. F., Revilla, E., Landscape and Urban Planning, 142, 198–211. https://doi.org/10.1016/j. … Alcaraz-Segura, D. (2020). Digital conservation in biosphere reserves: Earth landurbplan.2015.02.020 observations, social media, and nature’s cultural contributions to people. da Silva Lopes, T. H., Cadima Remoaldo, P. C. A., & Ribeiro, V. (2018). The use of photos Conservation Letters, 2020, 1–9. https://doi.org/10.1111/conl.12704 of the social networks in shaping a new tourist destination: Analysis of clusters in a Walden-Schreiner, C., Rossi, S. D., Barros, A., Pickering, C., & Leung, Y. F. (2018). Using GIS environment. In J. Rocha, & J. António Tenedório (Eds.), Spatial analysis, crowd-sourced photos to assess seasonal patterns of visitor use in mountain- modelling and planning. IntechOpen. https://doi.org/10.5772/intechopen.78598. protected areas. Ambio, 47(7), 781–793. https://doi.org/10.1007/s13280-018-1020- Small, N., Munday, M., & Durance, I. (2017). The challenge of valuing ecosystem services 4 that have no material benefits. Global Environmental Change, 44, 57–67. https://doi. Wilkins, V. M. (2004). Linking passive and active representation by gender: The case of org/10.1016/j.gloenvcha.2017.03.005 child support agencies. Journal of Public Administration Research and Theory, 16(1), Solombrino, O. (2018). Arcipelago Palestina—Territori e narrazioni digitali. MIMESIS. 87–102. https://doi.org/10.1093/jopart/mui023 Sonter, L. J., Watson, K. B., Wood, S. A., & Ricketts, T. H. (2016). Spatial and temporal Wood, S. A., Guerry, A. D., Silver, J. M., & Lacayo, M. (2013). Using social media to dynamics and value of nature-based recreation, estimated via social media. PLoS quantify nature-based tourism and recreation. Scientific Reports, 3. https://doi.org/ One, 11(9), 1–16. https://doi.org/10.1371/journal.pone.0162372 10.1038/srep02976 Stålhammar, S. (2021). Assessing People’s values of nature: Where is the link to van Zanten, B. T., Van Berkel, D. B., Meentemeyer, R. K., Smith, J. W., Tieskens, K. F., & sustainability transformations? Frontiers in Ecology and Evolution, 9, Article 624084. Verburg, P. H. (2016). Continental-scale quantification of landscape values using https://doi.org/10.3389/fevo.2021.624084 social media data. Proceedings of the National Academy of Sciences, 113(46), Stålhammar, S., & Thorén, H. (2019). Three perspectives on relational values of nature. 12974–12979. https://doi.org/10.1073/pnas.1614158113 Sustainability Science, 14(5), 1201–1212. https://doi.org/10.1007/s11625-019- 00718-4 18