Pathways to urban sustainability: Design perspectives on a data curation and visualization platform

Current computing methods addressing climate change employ open data platforms, interactive maps, and predictive modeling to support policy analysis, evaluate effects of policies, and support information access. These data exploration and analysis tools carry beautiful visions, including helping people better understand their cities, making communities’ challenges transparent, and facilitating systematic approaches to decision making for sustainability issues. They are also limited, and the reality they partake in turns out to be more nuanced. The tools that help us to get to know a city through ∗ Both authors contributed equally to this research.


INTRODUCTION
In 2022, our societies exceeded at least six of the nine safe planetary boundaries [79,94,102]. One of these gets the most attention. The 2023 synthesis report of the IPCC's Sixth Assessment Report (AR6) states that global warming of 1.1°C from 1950 levels has been reached in 2020 [45]. The unsustainable rates of consumption and production have caused previously unmatched impacts such as greater sea level rise since 3000 years, warmer temperatures than the last 125,000 years, and greater concentrations of carbon dioxide than the last 2 million years [45]. "There is a rapidly closing window of opportunity to secure a liveable and sustainable future for all. . . The choices and actions implemented in this decade will have impacts now and for thousands of years" [45, p. 25]. In order to limit warming to 1.5°C, decision-making for the mitigation and adaptation to climate change must be improved. Both top-down and bottom-up methods need to be applied in order to effectively create targeted actions that will bring us to the 1.5°C target.
IT has a role to play here in supporting analysis, modeling, prediction and assessment; in supporting us in handling the temporal, spatial and domain complexities involved in this planetary challenge [48,96]; and in coordinating for social change [6]. But its role is not so clear cut [14,15,53,54,57,70]. For example, machine learning has become a prominent tool used within climate change for forecasting, risk analysis, modeling impacts, urban planning, and informing policy [82]. However, recent studies have identified that the bias in predictive models can be attributed to a lack of auditing during data collection [9,47]. This leads to encoded bias within the predictions such as stereotypes, negative sentiment, racializing, or gendered bias. This raises additional issues of fairness, accountability, and ethics in data collection processes. Given that climate variability drastically affects all individuals and particularly affects vulnerable populations first and to a greater extent [31,37], we must scrutinize how predictive modeling used in decision-making processes can be made more fair, accountable, and transparent especially in its application within climate change.
Breaching planetary boundaries such as the climate is fundamentally unfair to those harmed by the effects, and sustainability is long recognized as an issue of justice. The concepts of environmental justice and just sustainabilities center the interlocking concerns of justice and the climate [3,4]. Just Sustainabilities have become an important guiding concern for urban policy making too [2,5]. This presents a challenge to design perspectives in HCI not just because it is technically complex but also due to its ethical complexity [15].
As both the public and private sectors are embracing computing and data driven methods to understand and address the challenges of just sustainabilities, these projects have also become more open to the public through open data portals [20,87], interactive maps [89] and public-facing reports on climate risks indicators [16] While these tools carry visions such as "encourage purposeful and easy engagement" [20] and "scaling innovative solutions to a wide range of local challenges," [89] the epistemology behind empowering communities through data and computing tools turns out to be much more nuanced. Scholars from urban planning, geography and computer science have argued that urban data and dashboards are not value-neutral systems that reflect objective facts about cities, but complex socio-technical systems that are battlegrounds for contested political, normative, and contextual ideas [51,61]. Although some of these data and visualization tools aim to empower local communities with easier and faster ways to interpret and manipulate data, they also prioritize a certain type of knowledge: scientific, formalized, encoded, and automatable, they center a knowledge that often appears more neutral and objective than it is [11,22]. As a result, they often marginalize local and indigenous ways of knowing generated from practice and experiences [51,62].
This paper attends to these tensions by demonstrating how design theories can orient the design process for computing systems used in both top-down decision-making and bottom-up community effort. The HCI and design community have a robust body of literature and frameworks for addressing complex sociotechnical challenges. In this paper, we specifically focus on four design approaches -Speculative Design, Social Design, Design Justice, Data Feminism -and explore how they can help us attend to the coexisting opportunities and challenges that arise from leveraging computational and data tools in addressing the climate crisis. We first position each approach conceptually and theoretically in addressing decision-making in climate change. Then we propose how each approach would inform and shape the design and development process of a tool meant to enable community advocacy and decision-making for climate change. Curbcut is an open-source map-based data platform, currently under development, aiming to support communities with sustainability data exploration and information sharing [19].
We unfold the challenges and opportunities in supporting climate change challenges by translating design theories into concrete actions that prioritize community values. These design theories therefore provide insight into how domain experts and community members can contribute to the decision-making process of mitigating and adapting to climate change. This alternate orientation of who is involved and what priorities are set will ultimately lend itself to improved outcomes for climate change decision-making.

RELATED WORK 2.1 Data, urban data, and their critiques
There is a rich history of data being generated, compounded, analyzed and presented to help answer questions about cities and their sustainability challenges. In the 1980s, official records of data started to be predominately processed and released in digital forms [50]. Since then, data generation has become more instant, finegrained and flexible across different scales and ranges of domains through the adoption of increasingly sophisticated techniques for both collecting and analyzing data [49]. Now, the heterogeneity and coverage of data continue to increase with sensors, surveys, social media, LiDar, crowdsourcing initiatives and administrative data all contributing to the generation of information about the physical and social environment that we live in [20,87].
Proponents of data-driven approaches to urban and sustainability governance believe that it is evidence-based, systematic, objective and generalizable -and that, by implication, anecdotal and localist ways of storytelling are not evidence [50]. Technical systems that generate these data supposedly take in reality directly and process it objectively into digital forms. These appear as politically and ideologically neutral facts. The resulting facts presented by urban and sustainability data guide research agenda and inform policymaking.
However, data is of course more nuanced than that. How data is categorized and organized is not a neutral process, but involves value, ethical and political judgment that has significant consequences for how it is interpreted and used [13]. Kitchin [50] links these systems of generating, collecting, and analyzing data to Foucault's concept of the 'dispositif,' where he argues that data for understanding cities are never objective but situated and used contextually to achieve certain goals. Therefore, sustainability and urban data, instead of "simply reflect[ing] the world," are forces that "actively produce" the world [51, p. 16] and shape how the world is "imagined" [55, p. 2].
When data is packaged into indicators, visualizations, maps and dashboards, there is more to be questioned and challenged. Mattern argues that there is always a governing entity who defines what data is important enough to be included in a dashboard, what is not important, how to present the information and what message to convey [61]. Dashboards tend to only present the cleanliness, "comprehensiveness and seamlessness" of data [61] and facilitate the illusion that all challenges can be captured in totality, visualized on screens, and all solutions can then be designed through interpreting and manipulating data [51,61]. On the other end of the user journey, when data is presented on a dashboard for users to comprehend, users are likely to have only a limited understanding of how the data were collected and analyzed. Needless to say, the existing dashboards are also doing very little to educate users about the context of data, politics of information visualization and knowledge production [61]. More generally, mapping, data visualization, and presentation is never objective -human perceptions, values and decisions are baked into data through data collection, analysis and communication [22]. When scholars who hold a critical stance towards urban data and dashboards recognize the limitations of data and computing tools, they do not deny their capabilities but stand with the tension to call for improved practices. Carefully designed and used urban data and visualization tools can serve as tools of empowering local activism as well as drivers for policy change. One example is the Environmental Justice Atlas [28].
The platform documents more than 3,800 cases of environmental conflicts across the world, contributed by local communities, as well as maps that visualize spatial data to tell stories about environmental injustice and activism that led to transformation. For example, "Pushed to the wasteland: Environmental racism against Roma communities in Central and South-Eastern Europe" shows the environmental injustice that Roma communities in Europe face through an interactive map that visualizes the percentage of Roma population and points of environmental risks and hazards [38]. The visualization presents a compelling story of how Roma settlements are being placed near mining and smelting complexes and landfills. The careful curation of historic cases, data visualization and rich descriptions of stories behind the numbers together shaped the platform into a database for grassroot environmental advocacy groups around the world to learn from and support each other.
Urban data and map visualizations can also act as crucial forces that push for policy change. For example, in recent research projects, short term rental listings' impacts on gentrification and rental gap were revealed through a series of map-based visualizations after large scale urban data collection and analysis [17,99,100]. These ongoing research projects led to a series of short term rental regulations and policy change across Canada [72,88]. As [51, p. 25] argues , "indicator and dashboard initiatives have much to offer city managers and citizens, but they need to make fully clear what is on offer. " Here, Kitchin advocates for projects that recognize their situatedness, relational effects, positionality, politics and who fully document their data lineage, metadata, error and uncertainty. In this way, these projects recognize other ways to see and understand the city [51]. [23] contend that data practitioners should "examine power", "challenge power", "elevate emotion" and "embrace pluralism" through an intersectional feminist approach. [62] emphasizes the importance of breaking out of the metaphor of seeing a city as a computer and highlights finding information that does not easily fit on a shelf or database through other forms of knowledge production such as site-based experience, participant observation and sensory engagement.

Data-driven Models for Climate Change
Data-driven visualizations and models of climate change have been used for many decades. In 1972, the Limits to Growth report used a global model to analyze key variables representing major global concerns including "... accelerating industrialization, rapid population growth, widespread malnutrition, depletion of nonrenewable resources, and a deteriorating environment" [65, p. 21]. At the time, the report generated a lot of criticism and debate specifically around its technical feasibility (model validity, data accuracy, assumptions, etc.) and its policy recommendations [85]. In their 30-year update, Meadows et al. further emphasized their original findings [64]. Both reports clearly state the need to use projections to make considerable policy changes and therefore aim to influence decision-making of experts and stakeholders on a global scale.
Similarly, the IPCC reports make use of IAMs to generate scenarios of future greenhouse gas emissions, air pollutants, land use, food systems, and more. The results from IPCC reports have had a significant role in informing policies on both a national and international level including the Paris Agreement, national climate pledges, and consequent adaptation and mitigation actions [78]. Since the first report in 1990, the IPCC has informed climate policymaking for over 30 years [12].
In recent years, the abundance of climate data and technological advancement has enabled the use of machine learning in both climate change mitigation and adaptation studies. This includes using ML to supplement policy analysis by evaluating policy effects and tradeoffs between policies [82]. However, the use of ML for decision-making poses challenges. Emerging machine learning research (MLR) has pinpointed the data underlying predictive models to be the largest contributor in introducing bias [77,84,86].
As a result, the importance of "data work" has been emphasized [84]. Data work in this context refers to performing data tasks, investigating data quality, applying frameworks around data practices, and the preparation of data prior to its use within a model to increase focus towards stewardship, quality, accountability, and transparency. A key aspect of data work is increased documentation. At present, MLR places the ethical onus on practitioners to reduce algorithmic bias. However, practitioners need robust, flexible, and easily integrated frameworks for widespread prioritization of data documentation to make systems more transparent and fair [40,43].
Evidently, the use of data-driven visualizations and models of climate change have had a substantial impact on expert decisionmaking for many decades. With advancements in computing from IAMs to ML, this trend will not decelerate. Rather, it is important to work through the challenges to ensure that future applications of modeling within climate change decision-making are fair and just.

DESIGN TENSIONS AND THEORIES
Designing in the age of the climate crisis mandates that we orient computing to operate within ecological limits, i.e., to prioritize human and environmental well-being over technology or unbounded materialistic growth [15,53,70,71]. In the context of computing for urban sustainability outlined above, this raises a set of interrelated challenges. Without claim for comprehensiveness, we will focus on: • The tension between lived experience and quantitative / formalized metrics and data (ways of knowing), • the tension between a tool as the object of design vs. the tool as a resource in a process (roles of technology), • The tension between expert driven design and policy making vs. community advocacy and engagement (governance), • The desire and ethical imperative to be inclusive vs. the difficulty of being genuinely inclusive (participation), and • The tension between prediction and impact assessment vs. the call to pluralize futures (futures) These tensions are certainly not new. The purpose here is not to address them exhaustively but rather to illustrate through these central themes and longstanding issues the contributions that specific design theories make to the urgent task of designing in the age of the climate crisis. To illustrate and address these tensions, we call on four design frameworks: social design, speculative design, feminist data practices, and design justice.
These four design theories are chosen to examine the tensions for several reasons. They all echo the need for critical reflection in the design process. This critical reflection often takes the form of understanding and incorporating worldviews that stem from differing social, political, and historical contexts. Furthermore, each of the theories provides unique motivations to prioritize in the design process (e.g.: collaboration, imagination, intersectionality, equity). While design theories such as inclusive design, action research, design for debate, and others would also provide key insights into the tensions, the four selected in this paper provide complementary perspectives on the design process. Each theory can be used to address one or more of the tensions but cannot provide the required lens for all tensions. Therefore, the complementarity of these four theories allows each of the tensions to be examined suitably.

Social Design
The history of social design (or designing for publics) can be traced back to the emergence of participatory design in the 1950s when labor laws reformed in Scandinavia and the U.S. to empower workers to participate in decision-making in the workplace [73]. Participatory design aimed at engaging users in a co-design process where the expert-driven top-down design process was contested. In a sense, the concept of social design is inseparable from a bottom-up approach to the climate crisis, as many of the early works that paved the foundation of social design explicitly discussed their motivation of addressing sustainability issues through design [58,59,76].
The value of social design is about engaging with communities to co-create new capacities to act collectively, with a focus on the process and commitments of design [34,59,69]. Therefore, the target of the design process shifts from focusing on the end material product, such as a computing system, to a process, in Le Dantec's words, that is "broader and is technology-agnostic" [56]. The objective of social design is not building an infrastructure, but to engage in the infrastructuring process that supports communities' future capacities in contending with the issues [56,93]. In other words, social design aims to empower communities to address future challenges beyond the current design process.
However, well-intentioned decision-making processes that aim to involve citizens can have shortcomings. For example, Mascarenhas and Scarce's study on consensus-based decision making to help resolve land management disputes in British Columbia discussed issues of participant competency and power like access to resources and negotiation experience [60]. This can have far-reaching implications as communities must be centered in order for the design process to be inclusive.

Speculative Design
Speculative design aims to augment the current market-driven purpose of design focussed on problem-solving to thinking beyond capitalistic uses and products to instead develop speculative designs to facilitate debate [26]. The purpose of speculative design is to spark discussion that generates new perspectives, inspires imaginations, and acts as a catalyst in redefining how we see design challenges and wicked problems. In this way, speculative design is a tool to see beyond fixed beliefs and worldviews to imagine new possible ways of being. Speculative design can take on many different forms, including art, fiction, exhibition, and workshops etc., where the participatory process of imagining and discussing alternative futures is valued more than the results or outputs [10,90,91].
Speculative thinking can be performed through the creation of fictional worlds. It acts as a method of reflection, critique, provocation, and inspiration. A method of developing fictional worlds is through thought experiments which are "constructions crafted from ideas expressed through design-that help us think about difficult issues" [26, p. 80]. One form of this is counterfactuals where a historical fact is changed to consider what would have happened to help in understanding the impact of key events. Thought experiments can also take the form of what-if scenarios which are similar to counterfactuals except they are generally focused on one major event and track the implications. Speculative design also coincides with design fiction, "the practice of creating tangible and evocative prototypes from possible near futures, to help discover and represent the consequences of decision making" [7]. As Eriksson and Pargman [29] argue, "breaking away from default modes of thinking about computing is difficult but possible", and necessary. Their counterfactual histories of the future of computing use design fiction to activate our ability to see the future in plural. This is key to building "a world where many worlds fit" [30].
Speculative design using fictional worlds is a flexible and imaginative way to consider the social, cultural, and ethical implications of science and technology within a society. It is praised as a method to unleash imagination, lower barriers for participation, create personal level connection and stimulate non-incremental change [90,101]. However, critiques of this design method reflect that there is a lack of clarity on actually conducting speculative design as a practical method [36]. Other common critiques state that speculative design tends to reinforce white, European perspectives without considering greater issues of inequality and may even underline existing power structures. For example, Revell argues that speculative design often generates dystopian futures in which social conditions are exacerbated into a dystopia to critique the current, European social standards rather than truly exploring alternate ways of being [95]. Some critics state that speculative design by generating alternate futures is not a critical method of design that can adequately be used to address fundamental political, economic, and environmental problems [92].
Developing fictional worlds in games is also a way of speculative thinking. For example, in the Climate Change Megagame, participants play roles like politician, business owner, or local resident in a city [97]. The designers provide the starting conditions, introduce counterfactuals, and participants co-create design futures. Players are presented with simulated climate change scenarios and they have to face choices resulting from impacts of the changed climate. The game is meant to create awareness, close the psychological distance related to climate change, and study the behaviour related to decision-making around climate change.

Feminist Data Practices
Data feminism is "a way of thinking about data, both their uses and their limits, that is informed by direct experience, by a commitment to action, and by intersectional feminist thought" [23, p. 10]. Key to the concept of feminist data practices are related understandings of intersectionality, power, and co-liberation. Along with being a movement for gender equality, feminism must also be intersectional by considering how race, class, (dis)ability, sexuality, age, religion, and other factors intersect to create unique identities that can lead to unique cases of privilege and oppression [23]. Complementary to this is understanding that power dynamics impact every experience and the power is unevenly held by straight, white, men from the Global North [23]. Building on past scholars from feminist science and technology studies, D'Ignazio and Klein argue for an alternative way of practicing data science, which pushes back against the mainstream patriarchal approaches of fantasizing problem-solving through data and argues for an experiential and pluralistic framework of working with data [23]. Co-liberation therefore strives for the understanding that power dynamics, privilege, and oppression ultimately are a hindrance not just to those who are oppressed but to everyone in society [23]. The book Data Feminism defines 7 principles to guide how feminism in data science work can be put into action. These include identifying and challenging power, valuing and prioritizing plural forms of knowledge, rethinking systems of oppression while empowering labour, and considering context for ethical analysis of situations [23].
The lack of consideration of intersectionality and issues of power within datasets used in models that facilitate decision-making lead to harmful outcomes for marginalized populations. Often, models seemingly created with the innocent motivation of improved efficiency and increased revenues are actually pervasive weapons that enable the unfair manipulation of a person's data [74]. This can be seen with models used for online targeting, determining employment prospects, gaming ranking systems, determining loan prospects, performing targeted policing, and many more [74]. Each of these examples demonstrate situations in which expert decisionmaking employs algorithms that intentionally or unintentionally cause harm and suffering.
Feminist data practices are also vital in decision-making related to climate change. The process of decision-making for any issue can be considered feminist if it "...challenges power by choice of subject matter, . . . by shifting the aesthetic and/or sensory registers of data communication, . . . or by challenging power by building participatory, inclusive processes of knowledge production" [23, p. 18]. Rosner [83] also touches on knowledge production by unveiling historical uses of feminized forms of craftwork to innovate engineering technologies such as the use of quilting to weave a memory core for a computer used for the Apollo space mission. Some of the questions she probes designers to consider is "Whose invisible work underpins your own? How might it inform your inquiry? . . . What representations feed a prevailing design narrative? Whom do they represent?" [83, p. 88-92]. Indigenous, local, and feminist knowledges can improve knowledge production around climate change by challenging the normative framework of climate change as techno-scientific [46]. Furthermore, the focus on the technoscientific perspective, albeit important, still only provides a partial viewpoint based on privileged designers and experts. This leads to marginalized worldviews to be excluded from decision-making that is based on techno-scientific models and its impacts [32]. For example, 90% of the victims of the 1990 flood in Bangladesh were women, partially due to the lack of warning information sent to them. Given that women and children are 14 times more likely to be fatally impacted by climate disasters than men, it is important for feminist voices to be included within decision-making as agents of change [35].

Design Justice
Design justice recognizes social and political influences and consequences of design choices and emphasizes community-led design to uplift local indigenous knowledge and challenge structural inequalities through an intersectional feminist lens [18]. The concept emerged and thrives with a growing community of diverse practitioners who aim to support more equitable, participatory and community-based design practice [1]. Ten shared principles for design justice aim to provide designers with guidance on rethinking design processes while using collaborative methods. The principles discuss the use of design to empower communities, centering the voices of impacted community members, making community members experts, and ensuring that design is a collaborative, iterative process [1]. Designing for justice includes considerations of historical context, situated context, and how social issues like class, race, gender, health and wellness impact people's experiences with technology and decision-making [33]. Costanza-Chock examines narratives, values, practices, sites, and pedagogies that design justice can help answer [18].
Successful applications of design justice can be seen in the examples of ICT infrastructure projects in the Detroit Community Technology Project (DCTP) and the Detroit Digital Justice Coalition [68]. These projects create communities involving actors from governments, policymakers, technologists, and more. Specifically, the DCTP runs data justice and equitable internet initiatives that nurture interpersonal relationships in order to engage and inform the public about internet-related concerns as well as rights around ownership and communication.

DESIGN THEORIES IN ACTION 4.1 The Action: Curbcut
Curbcut is a tool for urban sustainability data exploration developed by The McGill University Sustainability Systems Initiative. A version for Toronto is now in development in partnership with the Digital Curation Institute. It is a map-based data visualization and exploration platform, designed to generate and mobilize the knowledge to make cities more socially inclusive and less environmentally impactful, while improving the well-being of residents (figure 1). The tool aims to translate data into actionable knowledge to support policy and action by presenting multiscalar, temporal, accessible, understandable data for all. The current vision of Curbcut Toronto incorporates data from various perspectives, such as environment, transportation, demographics and economic etc., to afford systematic exploration of sustainability and equity from multiple facets. Users can not only visualize data in one dimension, as shown in figure 1, but also connect two dimensions for a correlation analysis. As figure 2 shows, users can visualize the bivariate relationship between number of trees per square kilometer and median household income. This flexibility for correlating different facets of just sustainability can potentially help uncover new insights relevant to policy making and support community activism efforts.
As designers of Curbcut, we hope to leverage our skills as HCI researchers to develop a tool capable of supporting system change and activism to address the climate crisis. At the same time, we are cautious of the pitfalls of and lessons learned from past failures of solutionist approaches [52,62]. We want Curbcut to realize its visions, be sustainable in the long term, and find its role in sustainability and civic works. At the current infancy stage of Curbcut, many questions and challenges lie ahead: how community values can be embedded in the platform and truly shape it (participation), how lived experience can be prioritized over quantitative metrics in the case of conflict (ways of knowing), as well as how the tool can transition from top-down researcher driven to community governance (governance). How may we address them?

Three Futures
We use design fiction to describe three potential futures. Imagine: 4.2.1 Future 1: Curbcut dies, community thrives. Five years after the launch of Curbcut, the initial excitement has waned, requests from the local community have not been responded to, and newly available data sets have not been made available on the platform.
Lily, a community activist leading a campaign on combating food insecurity, browses through information on food providers, she finds Curbcut and a visualization of accessibility to local food providers. But Lily is hesitant to use the information, as it does not provide sufficient context on when and how the data was collected, aggregated and visualized for her to assess its trustworthiness.
Lily closes Curbcut to find relevant and trustworthy information for the campaign elsewhere. After experiencing difficulties finding useful information online, Lily contacts other community groups in her network, asking if they have gathered or used relevant data. Luckily, with the help from another community group, Lily obtains the data she wanted on accessibility to local food providers. To obtain the information and comparative metrics specifically for her neighborhood, she needs to find a data savvy member from her team to clean and analyze the data. Through her volunteer network, Lily finds Megan, a data scientist who joins the team to crunch data and generate statistics and stories for the campaign. While policymakers and data scientists love Curbcut, community members have disengaged and stopped using it. They have lost trust in its top-down and expert driven approach. The information presented is overly complex and the visualizations incomprehensible. The visualizations lack context and some data seem to contradict the lived experience of community members. As an activist for combating food insecurity, Lily knows that what Curbcut presents about her neighborhood, suggesting an abundant amount of accessible food services, is not accurate. In fact, many traditional grocery stores, community gardens, and local businesses are struggling to remain open due to high rent increases over the past few years. However, given the apparent "comprehensiveness" of the platform, policymakers do not believe this kind of critique, claiming that Curbcut's comprehensive data aggregation allows them to understand local issues better than the residents.
Discouraged but resilient, Lily works together with other community groups to collect counter data [21,23] through telling stories of local food providers in her neighborhood to raise awareness on food security issues and challenges faced by local business. Just like any other activism works, large amount of time and effort has to be put in. Lily is up for the challenge. Community does not die easily.

Future 3:
Curbcut thrives, community thrives. Lily, a community activist combating food insecurity in her town, is leading a campaign. She accesses Curbcut, locates her neighborhood on the interactive map, and finds out that her neighborhood is at the 90th percentile of all neighborhoods in her city in terms of lacking accessible food resources. Lily reads through the documentation and learns about how the metrics were calculated, who contributed to the data collection and analysis and then tracked historical information for the past 5 years. She feel confident to bring this information directly to her campaign posters.
As an expert in local food resources, Lily realizes that a local food bank recently opened and is not reflected on Curbcut. She reports the new data to the platform's data team and Megan, a data volunteer, reads her request and incorporates Lily's data into the platform along with a thorough documentation.
In a later community planning meeting with residents and urban planners, Lily uses Curbcut to provide common understanding on food security in the neighborhood and provide background knowledge to facilitate conversation among people with different background understanding of the issues. Through exploring different data on Curbcut, such as green spaces, transportation, and housing, the group prioritizes opportunities and challenges for the next steps of their neighborhood's development plan.

Thrive or Die.
The most important question here is not if Curbcut dies or thrives but how the community thrives or dies and whether Curbcut helps it or harms it. We use design fiction to illustrate that a demise of Curbcut is not a tragic ending if community members have developed better ways to generate and share insights amongst each other that does not require such an artifact initiated by researchers. The tragic ending of the story is when Curbcut thrives as a tool driven by a solutionist, expert-driven and objectivist approach but marginalizes communities' engagement, culture, value and lived experiences.
At its current stage, Curbcut has the potential to better support community activism but carries deep tensions that are representative for larger struggles in design in the age of the climate crisis [70]. Central is the challenge of incorporating both conventional data analytics approaches and community-centered perspectives that prioritize lived experience. How we should approach the design and development process of Curbcut remains a meaningful question to be explored. As designers of Curbcut, we are currently in a unique situation: on one hand, we have a tool proven to be helpful for community groups in another city [24] at our doorstep, on the other hand, we have the opportunity to approach the design process carefully to recognize and address the different contexts, ways of operation, needs, and aspirations of communities in Toronto so that we can avoid the future of scenario 2 where a piece of technology marginalizes the local community.

Addressing the Tensions
In the next three sections, we explore how the five tensions (ways of knowing, governance, participation, roles of technology, and futures) can be addressed using four design approaches (social design, speculative design, feminist data practices, design justice) to inform the design and development process of Curbcut. We put these design frameworks into practice and explore: who should be involved, who should be considered as experts, what would the design process look like and what defines a successful design? By exploring the insights for one project, we hope to provide starting points for others too.

Ways of Knowing and Governance.
Applying data feminism and the design justice framework can allow Curbcut to serve as a tool that supports sustainability data exploration and the questioning of fundamental structural challenges faced by climate-impacted communities. These design frameworks emphasize social and political influences of design and prioritize values of the groups that are systematically disadvantaged and marginalized [18] given that the climate crisis and sustainability issues are always questions of social justice [3,15]. The questions and principles raised by these approaches provide an essential foundation and guidance for the design of a tool that attends to political and social challenges.
The data feminism principles and the design justice framework offers guidance for key questions: What should the target of the project be? How should we evaluate a successful project? Who should be involved in the process? In the context of addressing climate issues with Curbcut, these design methods also raise questions like: Who are the intersectional communities that are addressing climate advocacy? Whose voices are oppressed by the matrix of domination? Who is being privileged in the design process? Who will be privileged after the development of such a tool for accessing and visualizing sustainability data?
Examining and challenging power means finding and explaining forces of oppression in data projects and taking actions to counter the force [23]. This involves asking questions such as who is involved in preparing the data used in Curbcut, Who decides what data to include, who benefits from such a data platform and who is left out and neglected (governance). Design justice does not just advocate for the participation of those affected but requires them to be in charge (governance). Communities who are striving for climate advocacy work in their daily life should receive more attention and have a stronger voice in the design process than the researchers and developers (ways of knowing). As HCI researchers and designers, our primary role should be assisting community members and ensuring that their values guide the project.
Other concrete actions to challenge power include collecting counterdata [21], examining the positionality of all involved designers, questioning the context of data before using it, disclosing the context of how knowledge is made on the platform, providing detailed datasheets for datasets and providing user guides for all data included on the platform. Since data visualization is central to Curbcut, designers should document all the editorial choices that are made and acknowledge the framing effects, which impact how viewers interpret the graphics [23]. Furthermore, data visceralization should be explored in Curbcut (ways of knowing). Data visceralization liberates and leverages our emotion, such as fear, confusion, hopeful and solidarity through data visualizations to achieve a form of data maximalism [23]. Since Curbcut aims to address contested opinions and ideas around the climate crisis, designers must take a stand to express the urgency of actions through techniques such as data visceralization. Challenging existing power dynamics allows data practitioners and designers to embrace pluralism, attending to ways of knowing that are usually suppressed by traditional data work in service of cleanliness and control (ways of knowing) [23].

Participation and Roles of Technology.
Participation in design is a means to effect deep, transformative, socio-political change [34], which is necessary for addressing the climate crisis, a challenging, wicked, political and contested issue. Social design helps us to attend to the tension of the roles of technology and participation in addressing social challenges. In the social design process, Curbcut will be sitting at the periphery waiting to be examined and explored by the community. Whether it will be adopted or not depends on the articulation of issues, examination of attachments and the process of infrastructuring.
Social design engages with communities to co-create new capacities to act collectively (participation). The target of the design process is "technology-agnostic" [34,56,59,69]. In the context of designing Curbcut, the existence and plan to develop a data platform seems contradictory to the goal of social design. There is a risk that the designers are imposing a technological solution onto communities' sustainability issues. However, the important differentiation between a techno-solutionist process and a process of social design is where we place Curbcut in the design process and the design space. Instead of viewing Curbcut as a design outcome or solution to any issues, we can view it as a potential resource for communities and carefully explore its potential role in the already existing processes of communities' works.
Le Dantec's framework of social design consists of three key components: articulating issues, identifying attachments and the work of infrastructuring, which influence each other and together constitute a public [56]. A public is defined by Le Dantec as a configuration of individuals bound by common cause in confronting shared issues. In our case, the broader public could be any community or individual who hopes to address the climate crisis, with a focus on communities in Toronto. When articulating issues, it is noteworthy that issues are not fixed, but dynamic and evolving, and as the issues evolve the public might also change to include or exclude more people. The formation of issues does not aim to reach a precise definition, but a space to afford a plurality of views to create discourse around those views [56]. Then identifying attachments means finding multiples of relations across all actors, artifacts and institutions surrounding the articulated issues. In our context, this could be organizational structure within a climate advocacy group, the collaboration or competition between multiple groups, or the relationship between a volunteer and the group's information archival procedure etc. Ultimately, the work of infrastructuring aims to find neglected existing attachments and forms new attachments that aim to empower the public for addressing present and future issues. The key to infrastructuring is a focus on "design-for-future-use", which emphasizes enhancing communities' capabilities beyond the development and deployment of a single artifact [27,56]. The infrastructuring process of addressing the climate crisis from a bottom-up perspective could take the form of connecting different advocacy groups to work in solidarity, supporting climate data sharing within one organization, or updating current procedures of volunteer recruitment etc. Curbcut can come in to form potential attachments along with the existing social and technical network (roles of technology). The attachment formed by Curbcut might not be the most useful resource for addressing future issues, but it can serve as a potential possibility to be explored.
Social design opens a perspective of Curbcut as civic technology. As a recent study of civic tech suggests, the most exciting outcome of such a co-design process is not any tool in itself, but the fact that the process reconfigures the roles of democratic participation [63], contributing to a vivid community and wider social change.

Futures.
In the context of designing Curbcut, speculative design has the potential to serve as a first step to spark conversations and debates around sustainability advocacy work in different temporal scales and reflect on the role of technology in such processes.
Speculative design moves away from finding specific solutions to stimulate ideas, provoke discussions, questions, and debates that challenge the current dominant assumptions and status quo [26]. Potential speculation workshops can focus on a shorter term (5 years), where individuals and community groups who are concerned about the climate crisis can come together to imagine how ongoing effort in advocacy work can be augmented to make a larger impact that drives policy change or be suppressed and lead to negative impact (futures). A focus on the speculation workshop could be whether data tools like Curbcut can serve as an adequate support for communities ongoing effort and their potential harms to community works. When speculating the near term, more direct opportunities of support and action could potentially surface. Stretching the timeline to a longer extent, workshops can also focus on speculating a future world that lies 25 to 50 years ahead. Community members can collectively imagine many positive futures where their communities are more equitable, livable, and less environmentally impactful. Through the exercise of envisioning a utopia, we can indirectly reflect on the ways in which technology either supports or impedes the progress towards achieving this vision (futures).
Many previous works use speculative workshops to explore technology's (in)adequacy in addressing sustainability and justice issues [39,41,81]. Speculative workshops in the process of designing Curbcut could help us reflect on the opportunities and limitations of such a tool, but there could be so much more: as [103] put forward, speculative designs do not just construct artifacts but entire lifeworlds in which such artifacts exist and interact. This design approach offers the opportunity to zoom out from an artifact to consider the social, political and cultural context beyond the current worldview which can bring invaluable insights. However, all the critiques on speculative design are also valid in our context. We can have a seemingly participatory and gamified workshop, but how would this process actually make concrete impact to address sustainability issues in the long term? Will the speculative process be meaningful enough for us to address a pressing issue or diverging us away? A survey of speculative design papers remained "conflicted" towards speculative design, as it is hard to say how much it can offer in addressing the climate crisis [91]. We recognize that if speculative design is adopted in our context of developing Curbcut, we at least need to adhere to past scholars' provocation of inviting a diverse public and articulate and attend to our "unexamined assumptions. " [91]

DISCUSSION
The design theories discussed here provide insights, direction and momentum that orient design projects like ours. To illustrate briefly what that means in practice, we explore some steps we are taking.
To address the tension between lived experience and quantitative/formalized metrics and data (ways of knowing), we first explore the decentering of quantitative analytics in favor of storytelling. In its current form, Curbcut comes with some functionality for stories: narrative content enriched with multimedia, tagged with location and relevant metrics. Stories can be browsed and will be shown as corresponding context when a user analyzes specific categories. However, as it currently stands, this places narratives in second place. We will explore, with communities, what it would mean to flip the priorities and begin the users' exploration through stories, with data visualizations and analytics provided as context to illustrate the narratives. Second, we explore the possibilities offered by visceral visualizations inspired by feminist data practices.
To address the tension between a tool as the object of design vs. the tool as a resource in a process (roles of technology), we decenter the tool in community engagement and instead emphasize the process of infrastructuring, inspired by social design. If the tool needs to be put aside for the community to thrive, it should.
In these processes of engagement, decentering the tool also decenters the experts, but only to some degree. We address the tension between expert driven design and policy making vs. community advocacy and engagement (governance) by structuring our partnerships on the basis of design justice principles. For example, we have developed a project partnership with community organizations in which the community partners call the shots. In co-designing this partnership, community based research methods from other fields such as geography took priority, but both data feminism and design justice principles were welcomed and useful in shaping and articulating goals, methods, and roles. A crucial aspect of this is how such a partnership configures ownership and sovereignty over community produced data.
In projects like this, stakeholders with heterogeneous perspectives, diverse motivations and knowledges and uneven privileges meet. This always raises the risk that participation ultimately turns out to be a form of pseudo-participation in which participants are treated as data sources rather than decision makers [75]. We address the tension of participation via the structure of research proposals, and via a commitment to ongoing active reflection, following data feminist practices and critical systems thinking [15,25,66,67,98] Finally, the tension between unified prediction vs. pluralized futures is generative -in this uncertainty is where hope lies. It needs to remain open. Data visualization in Curbcut today is an account of the past. Two actions emerge. First, participatory speculative workshops will be a key component of the codesign of Curbcut. Second, we explore the integration of visceral visualizations of multiple possible futures, possibly backed by model-based simulations [48], as a generative technique for open-ended conversations about collective action.

Just Sustainabilities guide Computing within Limits
To envision a future within ecological, material, and energy limits, we must address social justice. In the conventional narrative of sustainable development, we are made to choose between the present or the future, or "between human welfare or ecological stability-an impossible choice that nobody wants to face. But when we understand how inequality works, suddenly the choice becomes much easier: between living in a more equitable society, on the one hand, and risking ecological catastrophe on the other. . . " [42, p. 144]. Ultimately, "justice is the antidote to the growth imperative-and key to solving the climate crisis" [42, p. 137]. This is why attention to justice, equity and participation has to be a central driving force in designing for sustainable futures.
To effectively address the climate crisis through design approaches in HCI, we must recognize the inextricably interlinked nature between justice and sustainability. This is the reason why design justice and data feminism are particularly appropriate to address the tension between expert-driven (top-down) and situated (bottomup) advocacy efforts around climate change. Scholars in sustainable HCI have recognized environmental sustainability as a social justice issue [53,70,71], and some sustainable HCI research explicitly focuses on environmental justice [44]. The framing of just sustainabilities, emerging from urban geography and environmental justice, makes this intersection explicit [3,4] and provides an apt foundation for the emerging framework of just sustainability design [15]. As Nardi suggests, ecological and environmental issues are a "predicament", a collection of multifaceted issues that require new assumptions and approaches. We support calls for a shift in emphasis in sustainable HCI [8,53,57] and emphasize the importance of an approach that addresses the ethical and social challenges at the intersection of sustainability and social justice [15].

Participation in Design
Participation, power and agency are recurring themes that underpin all four design approaches (social design, speculative design, data feminist practice and design justice). Ideally, through a participatory design process, affected communities are empowered to become co-designers, ensuring that the design reflects their values, needs and perspectives. In social design, designers immerse themselves with the public to together articulate issues and attachments in supporting the infrastructuring process. In speculative design, although participation is not a hard requirement, to achieve the design goal of sparking debates and discussion, only with the presence of the community can the process be complete. In data feminism and design justice, communities' mere participation is not enough. These two approaches argue for community ownership, emphasizing the importance of empowering marginalized communities to have more control in the design process.
While participatory design can appear to be an ideal process, it is important to remain cautious of limitations and critiques. Pseudoparticipation [75] occurs when community involvement is constrained by the limited forms of participation. In such cases, community members may have little agency or control over the final decision, and their marginalization may be disguised under the formality of participation. Even more concerning is when community participation is exploited solely as a source of data collection or as a way to legitimize decisions that have already been made [75,80]. As such, it is crucial to ensure that the participatory process is truly inclusive, empowering, and serves the needs of the community rather than merely tokenizing their involvement.

Complementing Design Theories
The design theories investigated above overlap in their aims and guiding principles while also offering distinct complementary concepts. For example, design justice and data feminism intersect in their principles. The feminist design principle to examine and challenge power coincides with the design justice principle to "... seek liberation from exploitative and oppressive systems" [1]. Both frameworks also emphasize the importance of valuing multiple forms of knowledge and placing those with lived experiences as equals amongst experts. There is also commonality between the discussions of labour, existing local practices, and experiential knowledge.
To be successful, social design requires not just community involvement but also community ownership in the design process. Both design justice and data feminism also advocate for this. However, both these frameworks go beyond the need for participation and also give importance to lived experiences by determining them as forms of expertise. Not only are community members participants in the data feminist and design justice approaches but the outcomes of the design process are centered around their articulated issues. Similarly, data feminism embraces pluralism by valuing that knowledge is not just expert-based but is also gathered from local, Indigenous, and pluralist perspectives.
Speculative design involves creating fictional worlds to explore and critique different perspectives, worldviews, and imagine possible futures. The process of speculative thinking can involve designing futures that incorporate the principles of data feminism and design justice including considering intersectionality, challenging power dynamics, acknowledging the social context of situations, recognizing the values embedded in designed things, and so on. Additionally, speculative design can occur with community involvement to increase inclusivity of who is involved in the design practice and to increase ownership of community members.

CONCLUSION
"We cannot return to any particular past, but nothing is stopping us from finding new ways to apply the wisdom of organizing human activity around community, simplicity, equality, and care, and directing design efforts toward these desiderata" [70, p. 14] In this spirit, this paper explores how design theories can orient design processes for computing within limits through policyfocused and community-centered efforts in addressing the climate crisis. We investigate the insights that four design approaches (Social Design, Speculative Design, Feminist Data Practices, Design Justice) provide for design in the context of climate change. We examined how their insights combine to address five key tensions: ways of knowing, roles of technology, governance, participation, and futures. We illustrated how we put these frameworks into action in the design of an open-source map-based data platform.