Making inequities visible: a call for participatory food system dashboards for SDG monitoring

1 Introduction

Cities are at the forefront of addressing the interconnected challenges of food insecurity, nutrition inequities, and food system transitions to more environmentally sound, healthier, and socially just futures (Hinrichs, 2014; Ilieva, 2016; Frantzeskaki et al., 2017; Cohen, 2022). The United Nations 2030 Agenda for Sustainable Development (UN General Assembly, 2015) calls for robust, science-based indicators to monitor progress toward the Sustainable Development Goals (SDGs), yet municipal governments often lack tailored tools to track food system performance at the level of disaggregation required to identify and address equity gaps (Simon et al., 2016). Recent work on SDG localization and the voluntary local review (VLR) process (e.g., Ortiz-Moya and Reggiani, 2023), which is a mechanism to localize the SDGs and the implementation of the 2030 agenda, further underscores the importance of local and regional governments in translating global goals into place-based monitoring systems, while also highlighting persistent challenges in developing indicators that are sufficiently granular, policy-relevant, and integrated across sectors.

Although global initiatives such as the Milan Urban Food Policy Pact (MUFPP) (Polman and Bazzan, 2023) and the Food Systems Dashboard (Fanzo et al., 2020; Schneider et al., 2023) have advanced the availability of national- and city-scale indicators, most existing dashboards remain limited in scope, focus primarily on national averages, and insufficiently integrate measures that reflect disparities by race, income, immigration status, or neighborhood context (Ilieva, 2017; Kitchin et al., 2015; Béné et al., 2024). Critical analyses of urban dashboards have shown that they tend to advance a narrowly conceived perspective that privileges quantitative measurement over other ways of knowing cities (Kitchin et al., 2015; Mattern, 2015), while food system dashboards in particular have largely been designed by panels of experts with minimal input from local stakeholders (Béné et al., 2024).

Moreover, while much of the literature on SDG indicators focuses on the data themselves and their potential to affect policies, programs, and governance by illustrating trends and providing outcomes data, less attention has been directed toward the processes by which metrics are selected, how data are used to raise questions and develop new causal theories, and how different stakeholders can co-produce indicators that reflect lived experiences and local contexts (Costanza-Chock, 2020; Béné et al., 2024). The dominant paradigm in urban food system monitoring has emphasized technical sophistication and data availability over community engagement and interpretive depth, as documented in prior critiques of urban data systems and dashboards. For instance, Kitchin et al. (2015) have argued that indicator and dashboard initiatives privilege scientific and technical knowledge while marginalizing knowledge derived from practice, deliberation, and lived experience. Mattern (2015, 2021) has traced the genealogy of urban dashboards to corporate and military control rooms, showing how managerial assumptions about performance measurement and urban legibility have been embedded in civic data tools. Most existing dashboards remain limited in their ability to capture the complex, place-based realities of food access and equity (Kitchin et al., 2016; Sarikaya et al., 2018). These platforms typically present static metrics that fail to incorporate community knowledge, provide little opportunity for user interaction, and rarely integrate qualitative insights that might explain discrepancies between official data and on-the-ground realities (D'Ignazio and Klein, 2020; Corburn, 2005).

Against this backdrop, this paper examines two prototype food equity dashboards developed in New York State: the Food Environment Equity Dashboard NYC (FEED-NYC) and the East End of Long Island Food Equity Dashboard. We present them together not as a controlled comparison of the same tool across two jurisdictions, but as linked applications of a shared participatory, equity-oriented dashboard approach in two contrasting food-policy contexts: a large, dense urban system and a mixed rural-periurban region. This paired case-study design allows us to examine which dashboard design principles, data integration strategies, and participatory processes are transferable across contexts, and which must be adapted to local institutional, geographic, and community conditions.

Furthermore, together, the dashboards are intended to function as local food system monitoring infrastructures: at the wider scale, they illustrate how municipal and regional indicators can complement SDG monitoring by grounding global goals in locally defined equity metrics; at the local scale, they respond to concrete governance and service-delivery needs, including municipal food policy planning in New York City and nonprofit coordination, advocacy, fundraising, and resource allocation on the East End of Long Island. Specifically, using these two prototype dashboard projects as case studies, this paper addresses three interrelated questions:

• How can municipal food system dashboards integrate traditional and novel data sources to monitor SDG-relevant indicators with an explicit equity lens?
• What processes and partnerships are needed to co-produce such dashboards in close and consistent collaboration with stakeholders?
• How do the resulting indicators from our pilot projects align with, extend, or diverge from the UN’s SDG indicator framework?

This paper is therefore positioned as an applied research and practice-oriented contribution that integrates empirical dashboard development with methodological reflection. Rather than presenting a purely evaluative or statistical analysis, we document how the two pilot dashboards were designed, the types of data and analytic strategies employed, and the insights that emerge from integrating traditional administrative datasets with novel and community-generated data sources. In doing so, we aim to make transparent both the “how” and the “why” of participatory food system dashboard development, while demonstrating their relevance for equity-oriented SDG monitoring at the municipal level. Additionally, by documenting the development process, data sources, and analytic strategies for these pilots, we also contribute to the literature on urban food systems metrics (Ilieva, 2017; Freudenberg et al., 2018; Specht et al., 2025; Rodman-Alvarez and Colasanti, 2019), participatory data visualization (Costanza-Chock, 2020; Dörk et al., 2020), and the role of cities in advancing SDG implementation (Simon et al., 2016; Zinkernagel et al., 2018; Ortiz-Moya and Reggiani, 2023). The paper offers actionable recommendations for policymakers and researchers seeking to design food system dashboards that capture both structural inequities and opportunities for more just, resilient, and sustainable urban food systems.

2 Materials and methods

2.1 Study design and approach

This study employs a mixed-methods, practice-based research design combining (1) integration and analysis of secondary administrative datasets, (2) development and analysis of novel data sources, and (3) qualitative data collection through stakeholder engagement. The methodological approach is designed to support both the construction of the dashboards and the interpretation of resulting patterns, with an emphasis on transparency and replicability. Rather than testing predefined hypotheses, the study takes an exploratory and applied approach to examining how diverse data sources can be integrated into interactive dashboards to surface patterns of food system inequities. The participatory orientation is informed by co-design approaches that treat affected stakeholders as contributors to problem framing, indicator selection, interpretation, and design rather than solely as end users (Sanders and Stappers, 2008; Costanza-Chock, 2020; Béné et al., 2024).

The analyses presented are primarily descriptive and comparative, intended to identify disparities, relationships, and potential areas for intervention rather than establish causal inference. We calculated counts, percentages, rates, ratios, geographic distributions, and ranges to identify disparities and potential service gaps across available spatial units. No hypothesis tests were conducted because the objective of the study was to develop and assess the utility of participatory dashboard prototypes, not to estimate causal effects.

The two dashboard cases—the Food Environment Equity Dashboard NYC (FEED-NYC) and the East End of Long Island Food Equity Dashboard—were selected because they are linked but distinct applications of a shared participatory dashboard approach within the New York State food-policy ecosystem. They are not intended to represent a controlled comparison of identical dashboard instruments across multiple cities. Rather, FEED-NYC served as a large, dense urban prototype focused on municipal administrative data, novel digital data sources, and citywide equity monitoring, while the East End dashboard adapted the same equity-oriented logic to a mixed rural-periurban region characterized by seasonal tourism, agricultural production, visible wealth, hidden poverty, and smaller nonprofit service networks. Presenting the cases together allows us to examine which design principles are transferable across contexts and which data sources, stakeholder processes, and policy linkages require local adaptation.

2.2 The New York metropolitan area context

The New York metropolitan area presents a unique context for advancing urban food system metrics due to its exceptionally rich though fragmented data ecosystem. New York City’s food policy landscape encompasses multiple data systems that were developed for different purposes and employ varying collection methodologies. For example, the Mayor’s Management Report (MMR) (NYC Mayor’s Office of Operations, 2025) provides annual operational performance metrics across city agencies, including indicators on school meal service and emergency food distribution. The Department of Health and Mental Hygiene’s Environment and Health Data Portal¹ offers environmental health indicators with neighborhood-level spatial disaggregation, including food environment assessments. Data2Go. NYC² serves as a community-level data platform integrating demographic, economic, and social indicators from federal and local sources. The Environmental Justice NYC Mapping Tool³ provides spatial analysis capabilities for examining cumulative environmental and social stressors, including food access dimensions. Many of these metrics are aggregated and organized in NYC Open Data, a publicly accessible repository of city data.

Despite this data richness, these systems remain somewhat disconnected due to their distinct institutional origins, varying geographic scales, different temporal coverage, and diverse intended audiences. Operational data collected for agency management purposes often lacks the neighborhood-level disaggregation needed for equity analysis. Environmental datasets may focus on regulatory compliance rather than community-experienced outcomes. Spatial data sources frequently employ incompatible geographic boundaries, making cross-system analysis challenging. This fragmentation means that critical aspects such as disparities in school meal participation, gaps between eligibility and enrollment in the federal Supplemental Nutrition Assistance Program (SNAP), spatial mismatches in emergency food access, and neighborhood-level food affordability are often addressed in isolation rather than through integrated, equity-focused monitoring systems.

Over the past two decades, among other global cities, New York City has been at the forefront of food policy and food metrics innovation in the US, leveraging legislation and data to track and improve food system outcomes. While there is an abundance of underutilized and under-integrated urban food system data, at times not even labeled as such (e.g., data in city management reports, city expense and capital budgets), notable gaps in publicly available data on key indicators such as neighborhood-level food costs, consumer experience and perception of proximate food outlets and choices (e.g., lived experiences and qualitative data), and many other non-food factors that shape local food environments (e.g., cost of housing, rent burden, cost of childcare, time poverty) are commonly missing from local and global dashboards.

2.3 Theoretical framework and research innovation

This research builds on several theoretical foundations that inform both dashboard design and the conceptualization of food system metrics for equity-oriented policy change. Drawing from critical planning theory, we recognize that data are not neutral but rather reflect power structures and institutional priorities embedded in their collection and presentation (Flyvbjerg, 1998; Kitchin, 2014; Leszczynski, 2016; Healey, 2020). Urban systems theory suggests that food access emerges from complex interactions between spatial, economic, social, and policy factors that require integrated rather than siloed analysis (Batty, 2007; Ericksen, 2008; Allen, 2012). Environmental justice frameworks emphasize the importance of disaggregated data that can reveal how food system inequities intersect with other forms of marginalization (Corburn, 2003; Alkon and Agyeman, 2011; Walker, 2012; D'Ignazio and Klein, 2020).

Our approach to dashboard development also draws from established theories of data visualization effectiveness and user interaction. Research in information visualization demonstrates that interactive dashboards have distinct advantages over static presentations by enabling users to synthesize information across multiple metrics, develop insights that may not be possible in other data presentation formats, and engage in exploratory analysis that supports evidence-based decision- making (Shneiderman, 2003; Liu and Stasko, 2010; Munzner, 2014; Sarikaya et al., 2018). Interactive dashboards serve as information management tools that combine multiple related visualizations into unified views while offering filtering, selection, and scenario modeling capabilities that engage users more directly in data analysis and the generation of new theories about relationships among variables in the food system. Compared to static visuals, interactive data visualizations have also been shown to induce greater thought processes, which in turn can lead to more positive attitudes about the data or the issue being presented.

We also incorporate framing theory to understand how variations in data presentation influence stakeholder perceptions and policy solutions. Framing involves selecting what is communicated and how information is presented, influencing issue salience and potential solutions (Entman, 1993; Bacchi, 2009; Hullman and Diakopoulos, 2011). Our pilot dashboard designs recognize that decisions about what data to present, how to visualize relationships, and which geographic scales to emphasize affect how users interpret and problematize food system challenges and envision interventions. The impetus for this work is not merely to fill existing data gaps, but to more fundamentally reconceptualize how existing indicators can be integrated into coherent frameworks that drive policy and systems change. Rather than treating food system metrics as isolated measures of program performance, we approach data visualization as a tool for surfacing structural relationships, challenging assumptions about food access, and enabling stakeholders to develop new causal theories about food system inequities. This reframing of data conceptualization represents a shift from technocratic approaches to dashboard development toward participatory knowledge production that centers community expertise and lived experience.

Our approach recognizes that the organization, presentation, and interactivity through which data are presented shapes not only what patterns and concerns become visible, but also what policy solutions appear feasible and desirable. By integrating quantitative administrative data with qualitative community insights, spatial analysis with temporal trends, and individual-level outcomes with structural determinants, the prototype dashboards test new possibilities for understanding food system dynamics and identifying intervention points for systemic change.

In response to the limitations of existing food system data visualizations, we developed two prototype, interactive data dashboards, one as part of a National Science Foundation (NSF)-funded Food Environment Equity Dashboard – NYC (FEED-NYC) planning project⁴ and, the second through a collaboration with community partners on Long Island’s East End region.⁵ The FEED-NYC prototype visualizes disaggregated indicators on school food, SNAP, emergency food access, and neighborhood food affordability for New York City. The East End Food Equity Dashboard adapts these concepts to a different geographic setting and type of local economy (i.e., centered on farming and tourism) and integrates poverty data visualization and mapping, participation in school meals, SNAP enrollment gaps, food pantry access, and grocery price variation. Both prototypes prioritize and are guided by the insights of community organizations, service providers, and government stakeholders from the start of the dashboard design process, combining administrative datasets (e.g., available through government open data portals) with complementary non-administrative data sources (e.g., web-scraped grocery prices, Google review sentiments, community narratives) to create equity-oriented, policy-relevant metrics. Together, these two cases represent complementary contexts: one large, dense urban system (New York City) and one mixed rural–peri-urban regional system (the East End of Long Island), allowing us to examine how dashboard approaches translate across different spatial, economic, and institutional settings (see Figures 1, 2).

Several features distinguish each project. In the East End dashboard, qualitative and narrative data were embedded alongside quantitative indicators, allowing stakeholders to challenge, contextualize, and explain findings, and to identify cases where official data did not reflect lived realities or the magnitude of need. In the FEED-NYC prototype, information collected for other purposes was used to convey information about food equity. For example, web-scraped grocery prices and Google reviews were used to assess affordability and food environment quality in ways that are not possible with existing public data. The dashboards were designed for interactivity, enabling users to model different scenarios, explore hypothetical changes, and view the data from multiple perspectives. Both prototypes sought to provoke new questions about the structural determinants of food access and to support ongoing dialogue among policymakers, advocates, and community members.

2.4 Analytic framework

To systematically explore and present existing efforts to further urban food system metrics, we developed and applied an analytic framework centered around several core questions: who creates the data, how they create and update it (e.g., with what level of stakeholder input), with what types of metrics (e.g., outcome vs. output indicators), covering what time horizon (e.g., longitudinal vs. cross-sectional), and for what audience (internal vs. external to government), see also Figure 3. An integral part of the analytic approach was also examining how data visualization strategies are implemented, including static versus interactive elements and the integration of narrative and qualitative data. For this paper, we also explored how dashboard metrics align with or diverge from existing SDG indicators through systematic comparison with the UN SDG indicator framework.

To illustrate how dashboard information could be used by decision-makers, we mapped selected indicators to existing policy and program areas as well as to potential decision-relevant applications. This mapping was designed to show how the dashboards organize data around actionable governance questions rather than presenting indicators as isolated metrics. The examples in Table 1 provide insights into how dashboard domains can inform both existing programs, such as school meals, SNAP outreach, and emergency food assistance, while also supporting broader policy and decision-making related to food affordability, transportation access, housing/rent burden, philanthropic grantmaking, and community-defined unmet need.

Table 1. Examples of policy and decision-relevance mapping of the dashboard domains and indicators.

2.5 Dashboard development processes

The analytic framework informed a broader participatory dashboard development process spanning project scoping, co-design, indicator and data-source identification, data integration and analysis, prototype development, stakeholder testing, iterative refinement, and dissemination. Figure 4 summarizes this shared process across the two dashboard projects and highlights the feedback loops through which stakeholder input informed revisions to earlier analytic and design decisions.

2.5.1 FEED-NYC dashboard development

The FEED-NYC dashboard project (2021–2023) aimed to co-design and test a prototype data, information, knowledge, and communication platform to address disparities in food insecurity in New York City. The research process included eight interrelated components (Figure 4): project scoping and partner identification; co-design and problem framing; identification of relevant indicators and administrative and complementary non-administrative data sources; assessment and application of methods to collect, clean, transform, integrate, and analyze the data; development of the dashboard prototype using appropriate modeling, visualization, and information-design techniques; stakeholder testing and qualitative feedback; iterative refinement; and preparation of three use cases illustrating how FEED-NYC could support government officials, service organizations, community-based advocacy groups, and other stakeholders in addressing food insecurity.

A structured co-design process was implemented based on collaborative methodology in which diverse stakeholders contributed to identifying problems and devising alternative solutions (Sanders and Stappers, 2008; Dörk et al., 2020). The implementation plan included three in-depth workshops to co-design novel food security indicators, identify analytical approaches to build the FEED-NYC prototype, and pilot its first iteration among community stakeholders. An introductory meeting discussed project goals, reviewed smart city concepts, identified and assessed existing data resources, and considered research questions and hypotheses.

Two co-design workshops with team members refined research questions related to food insecurity, identified indicators that stakeholders wished to prioritize, and began developing FEED-NYC by determining data needs, collection methods, and visualization objectives. Design thinking methodology (Brown, 2008; Dorst, 2011) was used to uncover, analyze, and visualize data in ways that revealed opportunities for policy intervention. A third workshop enabled team members, together with invited CUNY students and staff members from collaborating organizations, to review the prototype features and develop hypothetical use cases for city officials, service providers, and community leaders.

2.5.2 East End dashboard development

The East End dashboard development occurred in multiple phases beginning in February 2023, conducted in partnership with All for the East End (AFTEE) and with additional support from The New York Community Trust/Long Island Community Foundation (now The Trust-Long Island). AFTEE was a non-profit organization established to provide advocacy and funding support for the vast network of non-profit community organizations across the five East End towns of Long Island (East Hampton, Southampton, Shelter Island, Southold, and Riverhead). Originally founded in 2013, it gained prominence starting in 2020 through its “Feed the Need” campaign launched to quickly address the surging food insecurity and emerging human service needs in the region caused by the COVID-19 pandemic. Through this campaign, AFTEE raised and distributed grants to local food pantries and other non-profits.

The project used existing secondary data from government sources as well as manually compiled data (e.g., food prices for select grocery stores, community narratives via interviews), to create pilot dashboards that would uncover pockets of poverty, food insecurity, and inadequate access to healthy food that were otherwise invisible in aggregate data due to the high concentration of wealth in the region. Five thematic sets of visualizations were produced and implemented in Tableau Public: poverty on the East End of Long Island; participation in school meals compared to the share of economically disadvantaged students in the school district; differences between households living in poverty and households enrolled in SNAP; food pantry access and need considering neighborhood level food insecurity, vehicle ownership, and pantry service hours; and food retail prices and neighborhood food affordability.

During 2023–2024, the research team field-tested the dashboard prototype with local stakeholders through interviews and focus groups with community-based organizations to assess the value of the dashboard tool in their work and identify potential refinements. The team designed surveys and interview guides for primary data collection that elicited input from frontline providers on experience-based estimates of current unmet need, which was anecdotally known to be higher than represented by publicly available statistical data. An additional community insights tab of the dashboard portal was developed summarizing key input and narratives from stakeholders gathered through surveys and interviews, with relevant quotes from the qualitative data integrated into all other thematic dashboards. At the request of local stakeholders, an overview dashboard titled Town Profiles was created toward the end of the development process, presenting key food equity metrics and community narratives for each of the five East End towns.

2.6 Indicator selection, data sources, and integration

The indicators included in the two dashboard prototypes were not selected to constitute a validated composite index of food system performance. Rather, they were chosen iteratively based on four considerations: their relevance to policy and service-delivery questions prioritized by project stakeholders, their potential usefulness for local decision-making, the availability and geographic or temporal resolution of suitable data, and their grounding in existing food access and food systems research. Particular emphasis was placed on indicators capable of revealing disparities that may be obscured in aggregate measures, including differences in program participation, food affordability, and the spatial and temporal accessibility of emergency food services.

This approach is consistent with prior research demonstrating the value of spatial analysis for identifying mismatches between food-related need and the location of available resources (Bradley and Vitous, 2021), as well as recent reviews showing that measures of food pantry and food bank access benefit from considering multiple dimensions of accessibility, including distance, service availability, transportation, and other barriers rather than proximity alone (Safayet et al., 2024). Accordingly, the indicators were intended as decision-relevant, locally adaptable measures for exploratory and participatory monitoring rather than as a comprehensive or standardized measure of food system equity.

With this goal in mind, we integrated multiple categories of data sources across established administrative systems and novel data sources. Traditional administrative data formed the foundation of our analysis, including school meal participation data from the NYC Department of Education such as Daily Attendance by School and School Demographics Snapshots. SNAP enrollment information came from NYC Human Resources Administration data, SNAP Population by Borough statistics, and US Census Bureau American Community Survey (ACS) 5-year estimates. Food pantry data were gathered from Hunger Free America, Long Island Cares, and Plentiful reservation data collected by the NYC Food Assistance Collaborative. Demographic and economic indicators were drawn from the NYC Community Health Profiles, Feeding America 2020 food insecurity estimates, and New York State Education Department data (data.nysed.gov). City operational data included the Mayor’s Management Report (MMR), NYC Open Data, and the School Food Report.

We used Microsoft Excel to aggregate, clean, and organize the data. Integration across datasets was achieved by aligning geographic units (e.g., community districts, ZIP codes, census tracts) and standardizing variable definitions where possible. Quantitative datasets were combined through relational joins, while qualitative insights were integrated at the visualization stage through embedded narratives and annotations that contextualize observed patterns.

Novel and non-administrative data sources represented an important innovation. These included: web-scraped grocery prices from multiple supermarket chains in NYC and grocery store websites for Long Island; consumer sentiment analysis of Google reviews of food retail establishments to assess value perceptions and store experience (see Shen et al., 2018); mobility data derived from anonymized GPS data collected by SafeGraph (2019); community narratives collected through surveys and interviews; and food pantry utilization patterns analyzed using data from the Plentiful mobile app, providing insights into actual usage patterns and travel.

For the New York City food affordability analysis, grocery price data were scraped from the websites of a local grocery cooperative with locations throughout the city whose individual stores make independent purchasing and pricing decisions. A Python-based script was used to extract product names, prices, and UPC codes into a structured dataset (Hillen, 2019). The research team initially identified grocery item categories based on the USDA Thrifty Food Plan (TFP), which provides a model of a nutritious, practical, and cost-effective diet prepared at home for a reference family of four (U.S. Department of Agriculture, 2021). Prices for items within these categories were examined across store locations, including identification of cases in which an individual store’s item price was two or more standard deviations from the mean price for that item. To explore how food costs intersect with other dimensions of household affordability, store-level price patterns and identified outliers were also mapped in relation to neighborhood rent burden.

Developing a consistent citywide food basket was constrained by differences in products carried across independently operated store locations. Exact matches by UPC code were not available across all stores, making construction of a larger standardized basket unfeasible. The research team therefore narrowed the analysis to a smaller set of commonly available household staples, including items such as milk and eggs, that could be compared across a larger number of locations. The resulting indicator was intended as an exploratory measure of within-chain and neighborhood-level price variation rather than an estimate of total household grocery costs or a representative citywide food-price index. The grocer we focused on is also only one component of New York City’s diverse grocery retail environment and, thus, the analysis was designed to demonstrate the potential value of hyperlocal price monitoring rather than characterize the full range of prices available to consumers across the city.

For the East End of Long Island dashboard, the research team used convenience sampling to identify 17 supermarkets for which grocery prices were available online. We pre-selected a small illustrative basket of six common food items (i.e., cucumber, apples, whole-wheat bread, a dozen large eggs, 2% milk, and chicken breast) and recorded the lowest available in-store price for each item at sampled stores where all six items were available. The six-item basket was selected as a pragmatic subset of commonly purchased foods that could be compared across stores and it was not intended to represent a nutritionally complete diet or a comprehensive measure of household food expenditures. For each store, the prices of the six items were summed to create an illustrative basket cost, which was then mapped against village- and hamlet-level food insecurity estimates. Item-level price distributions were also visualized using box-and-whisker plots. This approach was intended to demonstrate the magnitude and spatial variation of food prices even within a narrow subset of goods and to assess the feasibility and policy relevance of incorporating locally specific food-price data into future food system monitoring.

All datasets were cleaned to standardize definitions of different variables and to ensure consistency when merging across datasets. These were all stored in a single repository located on a shared web-based platform, where version control was maintained. The team documented information on data sources, definitions, and any calculations and transformations. Across quantitative dashboard indicators, calculations were limited to descriptive measures appropriate to the available data and geographic scale. For school food, participation measures were calculated as the share of recorded meals or participating students relative to the relevant school enrollment or attendance denominator available in the source dataset. For NYC, the SNAP Program Access Index was calculated as the ratio of SNAP participants to the population below 125% of the federal poverty level, using ACS estimates and following the USDA definition. Data were summarized by community district and demographic subgroup where available. For the East End, the dashboard instead compared the number of households in poverty with the number of households reporting SNAP receipt. For emergency food access, pantry availability measures summarized counts of pantry sites and days or hours of operation, while pantry utilization and travel measures were calculated using client home ZIP codes and pantry ZIP codes. For food affordability, grocery basket measures summarized the prices of selected comparable items across stores or geographies, and rent-burden measures used the relevant ACS household denominator. These descriptive calculations were used to support dashboard visualization and exploratory comparison rather than statistical inference.

Because the dashboards were developed as prototypes using a combination of public, partner-provided, and platform-dependent data sources, replication should be understood as methodological rather than exact reproduction of every dataset. Core steps that can be replicated include identifying locally relevant policy questions with stakeholders, compiling administrative and demographic datasets, harmonizing geographic units, calculating descriptive indicators, integrating qualitative or community-generated data, and developing interactive visualizations organized around decision-relevant themes. Exact replication may be constrained where data are proprietary, partner-held, time-limited, or dependent on third-party platforms whose access rules, interfaces, or privacy requirements change over time, such as mobility data, app-based pantry-use data, Google reviews, or web-scraped grocery prices.

2.7 Participatory co-design and community data co-production

The prototype dashboards development process prioritized community engagement and stakeholder input through multiple mechanisms that recognized community organizations as both data sources and analytical partners. This approach aligns with broader “smart city” concepts that combine expanded urban data and technological capabilities with participatory governance, while emphasizing that technological innovation should be shaped by stakeholder participation and locally defined priorities (Meijer and Rodríguez Bolívar, 2016; Cardullo and Kitchin, 2019).

Stakeholder workshops formed a central component of our participatory approach. We conducted facilitated sessions using materials including beta versions of the sites and dashboards alongside brainstorming whiteboards with diverse participants including food policy advocates, food service providers, city officials, and researchers with expertise in data visualization and communication. We prepared the online brainstorming spaces in Miro – an online collaborative whiteboard platform which enabled us to discuss, plan, visualize, and document partner feedback on dashboard ideas in real time during project meetings and workshops. The workshops served not only to gather input on dashboard design, but also to identify community-held datasets and analytical insights that could enhance the dashboard’s utility.

Community organization partnerships revealed how groups providing food assistance services possess rich data resources, including client demographics, service utilization patterns, and qualitative insights about barriers to access, that often remain siloed due to limited analytical capacity. The research partnerships enabled these organizations to contribute their data while gaining enhanced analytical capabilities for understanding their service populations and advocacy needs. In New York City we engaged youth food advocates attending different public schools in providing input on data interpretation, enabling them to offer perspectives often missing from policy discussions and contributing student input to dashboard development.

To clarify how different stakeholders contributed to the dashboard development process, Table 2 summarizes the main stakeholder groups, the consultation or engagement formats used, and each group’s role in shaping the dashboards. These roles varied across the two projects and included identifying policy-relevant questions, contributing or interpreting data, validating findings, reviewing prototype dashboards, developing use cases, and connecting dashboard outputs to planning, advocacy, service delivery, and funding decisions.

Table 2. Stakeholder groups, engagement formats, and role in dashboard development.

Community reporting sessions were conducted with East End community-based organizations and nonprofit leaders to validate findings and gather additional insights. For the primary data collection, the research team developed a 10-item online survey, administered via Qualtrics, and reached out to representatives from thirty seven frontline service organizations operating across the five towns of the East End of Long Island. Twenty organizations completed the survey between November 2023 and January 2024, yielding a response rate of 54.1%. Participants were recruited purposively to capture a range of organizations serving different populations across the region. Of the twenty survey respondents, representatives from ten organizations were subsequently interviewed during the same period. These approximately one-hour interviews included organizations serving all five East End towns and were designed to provide a deeper understanding of their clientele, perceived changes in food insecurity and unmet need, factors contributing to food insecurity, and organizations’ experiences providing services in recent years. Survey data were analyzed using descriptive statistics, while interview transcripts were analyzed using thematic coding, with recurring themes identified iteratively to interpret patterns in perceived unmet need, service gaps, and barriers to food access. All qualitative coding and interpretation were conducted by members of the research team; generative AI tools were not used to code, analyze, or interpret survey responses or interview data.

This community engagement process was designed to capture experience-based estimates of current unmet need, which stakeholders indicated was higher than what was represented in publicly available statistical data. The dashboards also incorporated interactive and crowdsourcing elements through user feedback mechanisms and opportunities for community members to contribute additional data points, including narrative accounts of food access experiences and updates on changes in local food environments. This approach recognized dashboard users as active contributors to ongoing data collection rather than passive consumers of information.

2.8 Dashboard and website implementation

We considered various factors in the selection of the visualization platform to build the prototype dashboards. Preliminary sections of the FEED-NYC prototype were initially built in a Python-based environment. While Python-based dashboards offer greater customization and analytic capabilities, they also require dedicated coding expertise to build and maintain the dashboard. Recognizing the need for multiple team members to contribute simultaneously to dashboard development, the team transitioned to Tableau, which had a lower learning curve and offered data visualization capabilities sufficient for our needs. The team developed the dashboard in Tableau Desktop (2022.4), where each visualization (e.g., map, scatterplot, bar chart) was created as an individual worksheet. Multiple worksheets were then compiled into dashboards based on common themes/topic areas. Interactive features (filters, sliders, etc.), navigation buttons, and explanatory text, titles, and legends were added. These working versions were published to Tableau Public where they were accessible to partners and stakeholders. The dashboards were iteratively refined through stakeholder testing and feedback, improving both usability and clarity of content. Although we initially created the dashboards in Tableau Desktop, we could have opted to directly develop them in Tableau Public, since the end products were public-facing dashboards consisting of non-confidential data.

We created two public websites developed through Google Sites, a free web building platform by Google to house all the completed dashboards in one location, allowing us to share the dashboards broadly for each project with a single website address. The home page of each website provided an overview of the project, and thematic pages corresponding to the key food system domains (e.g., school meals, SNAP participation, food affordability) were created (see also Supplementary Materials, parts S1 and S2). Dashboards were inserted into the appropriate page of the website using the embed code copied from each Tableau Public dashboard. Accompanying text was added to each page to provide context for the underlying issue and help with user navigation of the dashboards.

2.9 Evaluation methodology

We used a multi-level evaluation framework that moves beyond simple usage metrics (such as page views or user interactions) to assess more meaningful measures of impact. For the FEED-NYC dashboard specifically, project evaluation methods included periodic feedback from team members about the workshops and structured interviews with a sample of participants. The evaluation approach examined both the technical aspects of the equity-focused data dashboard and the social processes involved in creating and using the dashboard for operations, service provision, and policy development.

Because both dashboards were developed and tested as prototypes, the evaluation did not seek to measure direct causal effects on enacted policies, budgets, or program outcomes. Instead, it focused on intermediate indicators of policy relevance, including whether the dashboard development process helped stakeholders clarify priority questions, interpret data across sectors, identify potential service gaps, and assess how the tools could support planning, advocacy, service provision, fundraising, and future policy discussions.

Three detailed use case scenarios were developed as written and illustrated narratives explaining how the dashboard would be used by different stakeholder groups: city officials, service providers, and community leaders. These use cases were shared with a wider network of stakeholders for feedback about potential applications of FEED-NYC to policy development, service provision, and program management. The use cases enabled stakeholders to reflect on whether and to what extent data collection and analysis, and visualizations of sociotechnical systems like the food system, function as heuristic guides to shape strategies and perceived solutions. For the East End dashboard, evaluation focused on stakeholder feedback gathered through the surveys and interviews described above, including perceived usefulness, clarity, and potential applications for planning, advocacy, service delivery, and fundraising.

3 Results

3.1 Dashboard profiles and key metrics

Both dashboards integrated traditional administrative datasets with novel data sources to provide tailored views of local food systems. Researchers combined more than 15 distinct datasets across both dashboards. The Long Island dashboard integrated quantitative metrics with qualitative community insights, while both dashboard prototypes enabled geographic visualization at multiple scales including neighborhood, census tract, and community district levels. The team also incorporated temporal dimensions to show trends over time, providing users with both current snapshots and historical context for some of the metrics. The results presented below reflect patterns identified through the integration of multiple datasets and qualitative inputs, as described in the Methods section.

The NYC dashboard featured disaggregated indicators across all five boroughs with community-district and demographic characteristic breakdowns. It integrated school food, SNAP, emergency food access, and neighborhood food affordability data through interactive visualization elements that enabled user exploration. The East End of Long Island Dashboard adapted these concepts for a focus on five towns: East Hampton, Riverhead, Shelter Island, Southampton, and Southold. This dashboard integrated the farming and tourism economic context specific to the region, incorporated poverty data visualization and mapping, addressed rural-specific food access challenges (e.g., including access to public transit), and pioneered community narrative integration approaches.

The dashboard development utilized a systematic approach with the thematic components implemented in Tableau Public, creating a coherent website presentation intended to facilitate navigation across dashboard themes. Field testing with local stakeholders provided validation and refinement opportunities, while primary data collection through surveys and interviews with frontline service providers captured experience-based estimates of unmet need that exceeded official statistical representations. In the East End of Long Island dashboard pilot, the integration of qualitative stakeholder input through a dedicated community insights tab and embedded quotes throughout thematic dashboards represented a methodological advance in food system monitoring.

The figures presented in this paper were selected to illustrate the main methodological and analytic contributions of the dashboard prototypes rather than to provide an exhaustive inventory of all available views. Each figure highlights a distinct function of participatory food system dashboards: multi-scale geographic filtering and overlay; integration of quantitative indicators with qualitative narratives; visualization of service availability and potential need; town-level synthesis for local decision-making; and demographic/geographic exploration of public benefit participation. The Results sections below therefore use selected dashboard views as illustrative examples while the full compendium of dashboard panels, including 16 FEED-NYC panels and 10 East End Food Equity Dashboard panels, is provided in the Supplementary Materials files accompanying the paper and on each dashboard website.

3.2 Emerging questions from the process: equity dimensions and access insights

The dashboards revealed several important findings about food system equity across multiple domains. Here we briefly highlight four of them. These four findings were selected based on their recurrence across both dashboards and their relevance to key dimensions of food system equity (access, affordability, utilization, and program participation).

3.2.1 School meal participation disparities

Despite New York City’s adoption of universal free school meals for all students, regardless of household income, nearly a decade ago, participation in the school breakfast and lunch programs varies notably across the city and among schools within the same neighborhoods, even when accounting for school-level rates of student poverty (i.e., schools with high proportions of students from low-income families do not necessarily have higher school meal participation). This suggests the need to better understand the causes of variation in school food participation and opportunities to advocate for improvements to the school food program in low-participation schools.

3.2.2 SNAP enrollment vs. eligibility gaps

Levels of participation in the SNAP among households estimated to be eligible for SNAP based on income thresholds derived from ACS data are uneven across New York City community districts. Estimates based on the Program Access Index (PAI) indicate that participation among eligible populations varies widely across districts, with some communities enrolling fewer than half of those estimated to be eligible. This finding suggests opportunities for more targeted outreach by local government agencies and service providers in under-enrolled neighborhoods.

3.2.3 Emergency food access patterns

Within New York City’s extensive network of emergency food pantries and the sustained efforts of providers to meet community needs, data on food pantry use show that many individuals travel outside of their own neighborhoods (ZIP codes) to get emergency food, even when there are one or more pantries in closer proximity to their homes. While food preferences (e.g., type, quality, quantity) are likely a key driver, part of this can also be explained by the limited service hours during which the pantries are open to individuals.

For the East End, specific challenges emerged around the limited number and hours of food pantries as well as access to adequate transportation (see also Figure 5). There were 32 food pantries on the East End of Long Island, but their uneven spatial distribution and service hours could pose potential access challenges for those struggling with food insecurity. Only 11 pantries were open more than 3 days a week and only 10 were open after 5 p.m., making them less accessible to those who work during regular business hours.

3.2.4 Neighborhood food affordability and price disparities

The prices of identical food items varied substantially at stores across a large chain of independent supermarkets located throughout New York City. In some neighborhoods with high rates of food insecurity, both rent burdens and food prices were high. Wide variation in prices was also observed when looking across various grocery chains, both in New York City and the East End.

3.3 Traditional vs. novel data integration

The two prototype food equity dashboard projects illustrate the value of integrating non-traditional data sources with traditional administrative datasets. Our descriptive comparisons across datasets indicated distinct advantages of combining these approaches, with traditional data utilization providing foundational demographic and programmatic data through government open data portals, offering reliable baseline metrics for program participation, enabling longitudinal analysis and policy compliance monitoring, and supporting cross-jurisdictional comparisons through standardized collection methodologies.

Novel data source pilot data collections and visualizations in the dashboards proved particularly valuable in several domains. Through an analysis of publicly available price information scraped from the internet, we identified substantial variations in prices for essential grocery items within a supermarket chain within and among different neighborhoods. Consumer sentiment analysis conducted through content analysis of Google reviews of supermarkets revealed how sentiments of price and quality varied according to the supermarket chain and by location. Mobility pattern analysis using anonymized GPS data gathered from smartphone locations enabled us to analyze shopping travel patterns, specifically examining home census block groups of shoppers and the food retailers at which they shop for groceries. Service utilization analysis using data collected by a mobile app for clients of food pantries allowed us to analyze which food pantries clients visit and how far they must travel from home to get to their preferred food pantry.

3.4 Visualization approaches: community insights and narrative integration

A distinctive feature of the Long Island dashboard was the integration of qualitative community insights alongside quantitative data. Integration occurred at both the analytical and presentation levels: qualitative narratives informed the interpretation of quantitative patterns and were also embedded directly within dashboard views alongside quantitative indicators, allowing users to consider statistical patterns in light of lived experiences and stakeholder perspectives.

The East End dashboard team surveyed key stakeholders and included their voices to provide context for understanding quantitative trends. A dedicated community insights tab was developed that summarized key input from stakeholders gathered through surveys and interviews, with relevant quotes from the qualitative data integrated into all thematic dashboards. The insights were also integrated in summary tabs for each of the five East End towns (see also Figure 6). This approach addressed the gap between official statistical representations and frontline provider estimates of unmet need, which were consistently higher than what appeared in publicly available data.

The aims of the qualitative surveys and interviews were to gain a better understanding of the organizations’ clientele, factors contributing to food insecurity, and the experience these organizations have faced in recent years. For instance, narrative data integration proved essential for interpretation of trends in the region, with survey respondents highlighting how “The (food access) need continues to grow at an alarming rate” (Survey respondent from a food pantry in Southold) and describing how “Costs for basic needs: housing, shelter, clothing, transportation have increased the level of uncertainty/insecurity of living on the Long Island far reaching into higher income residents” (Survey respondent from a food pantry in Southold).

Others further underscored the impact of housing costs on household budgets in the community: “The cost of rent in the last year or so has basically doubled in the Riverhead and Southampton townships […]. We do not have any low-income housing” (Director of a food pantry in Southampton). These and many other narratives elements helped explain patterns visible in the quantitative data and provided explanations for why individuals might travel longer distances to access food assistance or why certain grocery stores might be perceived differently despite similar objective characteristics.

3.5 Interactive features and cross-system analysis

The integration of static and interactive elements distinguished both pilot dashboards from traditional reporting tools and enabled exploratory analysis of potential policy interventions. Geographic and demographic filtering capabilities allowed users to examine data at different spatial scales and focus on specific neighborhoods or regions while disaggregating by race, income, and other characteristics where available. Multi-scale spatial analysis was implemented at neighborhood, census tract, and community district levels, enabling identification of patterns that might be obscured at any single geographic scale. Temporal trend visualization showed changes over time in some of the key indicators, providing users with both current conditions and historical context for understanding food system dynamics. For example, in the FEED-NYC dashboard users can view SNAP Program Access Index (PAI) data for 2019–2021 to visualize how access changed over time across different demographic groups, offering both current insights and historical context for identifying potential gaps in enrollment or underutilization of government assistance programs (see also Figure 7).

The dashboards laid the groundwork for analytical capabilities that could examine a wider and deeper set of potential factors underlying food inequities and better inform targeted public policy interventions. Specifically, cross-system analysis capabilities (e.g., food, housing, transportation) and overlay functions enabled users to examine relationships between different aspects of the food system, such as comparing food costs with poverty rates, travel patterns with service availability, and demographic characteristics with program participation rates. This integrated analytical approach revealed patterns and relationships that would not be apparent when examining individual metrics in isolation, supporting a more comprehensive understanding of food system dynamics and more strategic approaches to intervention design.

For instance, food pantry operating hours visualizations and travel patterns to and from home ZIP codes enabled dashboard users to focus on and explore important relationships between service availability, potential need, and utilization patterns. SNAP dashboard visualizations facilitated reflections regarding potential policies for targeted outreach in neighborhoods under-enrolled in SNAP by examining gaps between estimated eligibility and current participation rates, suggesting specific areas for intervention. Rent burden layers allowed users to raise questions about how policies like rent freezes or affordable housing interventions might affect household food budgets and the importance of examining the relationships between housing costs, food prices, and estimated food insecurity levels.

3.6 Alignment with SDGs

The SDGs were not embedded as a labeling or navigational framework in the original dashboard prototypes. Rather, for this article, we retrospectively mapped dashboard indicators to relevant SDGs and targets to examine areas of alignment and dimensions of local food-system equity that are not fully captured by existing international frameworks. This distinction reflects the dashboards’ original co-design process, which was organized around locally defined food-equity and policy questions rather than the SDGs themselves.

3.6.1 Areas of alignment

Direct alignment with several SDGs was evident across key areas (see also Table 3). The data corresponded with SDG 2 (Zero Hunger) through indicators on food insecurity, access to nutritious food, and participation in food assistance programs, with neighborhood-level disaggregation that allows identification of persistent food insecurity hotspots within local communities. SDG 10 (Reduced Inequalities) was reflected in the emphasis on equity and disaggregated data that could support monitoring of inequality reduction efforts, particularly through analyses of differential access patterns such as SNAP enrollment by race/ethnicity, income, age, and nativity. SDG 11 (Sustainable Cities and Communities) was connected through indicators relevant to urban planning and policy development on food system sustainability, including transportation factors and neighborhood- or school district-level availability and utilization of food assistance programs. Finally, SDG 17 (Partnerships for the Goals) was exemplified in the collaborative development process of the pilot dashboards, which brought together government, community, and academic partners and could serve as a model for multi-stakeholder collaboration in advancing SDG implementation.

Table 3. Overview of prototype dashboards themes, example indicators, and related sustainable development goals (SDGs), showing how local food system metrics align with and have the potential to deepen monitoring on multiple global goals within the 2030 Agenda framework.

3.6.2 Gaps and opportunities for extending the SDG frameworks

The dashboards also revealed several dimensions underrepresented in the current SDG indicator framework (UN General Assembly, 2017), offering valuable ways in which the SDG set of targets and indicators could be further detailed and grounded at the local level (see also Table 4). Narrative and qualitative indicators proved important for understanding quantitative patterns and identifying intervention priorities, yet current SDG frameworks provide limited guidance for incorporating qualitative data and indicators systematically. The ability to include and overlay non-food systems and sociodemographic characteristics of the population emerged as critical. While SDG 2 focuses on hunger outcomes, the dashboards and the findings presented here suggest the importance of upstream factors (Cohen, 2019; Cohen and Ilieva, 2021) like housing costs, transportation access, and work schedules that determine food access but fall outside traditional food security and food access monitoring.

Table 4. Locally relevant indicators from the FEED-NYC dashboard prototype that extend SDG frameworks by capturing city-level dimensions of food access, equity, and system adequacy.

Further, community-defined metrics revealed that communities prioritize indicators such as cultural appropriateness of food options, dignity of service delivery, and community ownership of food resources that are not fully reflected in current SDG frameworks and targets. A multi-pronged approach to equity analysis showed that current SDG monitoring often treats different forms of inequality separately, while the pilot dashboards developed for two starkly different geographic areas in the State of New York – New York City and the East End of Long Island – demonstrated the value of analyzing how multiple axes of marginalization interact to shape food access patterns. These findings suggest opportunities for expanding SDG frameworks to incorporate more nuanced approaches to measuring progress on urban food system transformation.

4 Discussion

4.1 Significance and implications: advancing local and regional food system metrics for SDG monitoring

Our findings demonstrate that cities that already have the data resources and institutional capacity to develop local food system monitoring tools can help advance SDG implementation through local data infrastructure and lived experience insights. The integration of traditional administrative data with novel sources creates opportunities to develop a more nuanced understanding of food system performance and to better inform targeted policy decisions. The participatory development process we used in developing the two prototype dashboards proved essential for ensuring that the dashboards reflected stakeholder priorities and incorporated knowledge that would otherwise be invisible in official data systems. Community organizations brought critical insights about data interpretation, helped identify gaps between official metrics and lived experiences, and provided ongoing validation of findings.

Additionally, the two pilot dashboards we prototyped demonstrate how local and regional food system monitoring can advance data reports and visualizations beyond static metrics to incorporate community knowledge, provide interactive exploration capabilities, and support evidence-based policymaking. Cities and regions are well-positioned to lead both in monitoring progress toward and advancing implementation of UN SDGs through next-generation food system metrics that prioritize equity, community engagement, and systemic understanding of food access determinants. Although the dashboard prototypes were not initially developed with the SDGs as an organizing framework, future iterations could incorporate a metadata layer that maps each indicator to relevant SDG targets, enabling municipal-level tracking of progress toward the 2030 Agenda and beyond. This aligns with the growing literature on SDG localization and voluntary local reviews, which emphasizes that local monitoring processes can help connect global sustainability frameworks to municipal policy integration, priority setting, and accountability, but require locally meaningful indicators and institutional mechanisms for translating data into action. More broadly, integrating such an SDG layer across local and regional food system dashboards could strengthen the feedback loop between local data infrastructures and global sustainability monitoring efforts, providing the UN and partners with grounded, place-based insights.

Considering the broader literature context, this work builds on research such as Sarah Williams’ “Data Action” (Williams, 2022), Catherine D’Ignazio and Lauren F. Klein’s “Data Feminism” (D'Ignazio and Klein, 2020), and Sasha Costanza-Chock’s “Design Justice” (Costanza-Chock, 2020), which collectively argue that data visualization and design processes must center the knowledge and participation of marginalized communities rather than reproducing existing power structures. It also extends critical analyses of urban dashboards (Kitchin et al., 2015; Kitchin et al., 2016; Mattern, 2015) into the domain of food system monitoring, contributing to the literature on urban food systems metrics and the role of cities in advancing SDG implementation. Recent work by Béné et al. (2024) has similarly argued that co-construction is the missing element in most food system dashboard initiatives, finding that the non-participatory nature of existing tools limits their relevance and uptake by decision-makers.

4.2 Lessons learned: methodological insights

Several elements of methodological novelty were tested through this work which, if developed further and beyond the prototyping phase, can help meaningfully advance both food system monitoring and participatory data visualization practice. The ability to examine data at multiple geographic scales, from census block to community district, enabled identification of patterns that might be obscured at any single level of analysis, revealing the importance of multi-scale approaches to food system monitoring. Data integration innovations demonstrated that existing data collected for other purposes, such as mobile app usage, web reviews, and mobility patterns, proved valuable for understanding food system dynamics when analyzed through an equity lens. However, some of the novel data sources employed, including web scraping, GPS data, and app usage, may not be consistently available or may change due to privacy regulations or platform modifications, raising questions about long-term sustainability.

The combination of quantitative metrics with qualitative community insights provided richer analytical frameworks than either approach alone, enabling more nuanced interpretation of food access patterns and barriers. Participatory development insights revealed that community organizations possess rich data resources that often remain siloed due to limited analytical capacity, suggesting opportunities for enhanced collaboration and capacity building. The co-production of data through community partnerships enhanced both data quality and stakeholder engagement, creating more accurate and actionable monitoring tools. Integration of lived experience data proved essential for understanding quantitative patterns and identifying intervention priorities, highlighting the limitations of purely technical approaches to indicator development. Despite innovations in data collection, large gaps remain in understanding lived experiences of food access, particularly for undocumented immigrants and other vulnerable populations who may be underrepresented in administrative data.

As university-based research partners, we occupied dual roles in these projects as both conveners and facilitators of the participatory process and as technical research partners responsible for data compilation, integration, analysis, visualization, and synthesis. We therefore were not neutral observers of the dashboard development process: our disciplinary perspectives, prior work on urban food policy and food-system metrics, and technical capacities influenced how problems were framed, which data sources and indicators were considered feasible and relevant, and how information was analyzed and visualized. At the same time, these decisions were not treated as fixed. Through co-design workshops, partner meetings, prototype review, interviews, surveys, and reporting sessions, we sought to keep research questions, indicator choices, interpretations, and visualization decisions open to stakeholder critique and iterative revision. This dual role underscores an important methodological tension in participatory dashboard development: university partners can contribute technical capacity and help connect otherwise fragmented data, while also exercising influence over which questions, forms of evidence, and representations become visible. Reflexive and iterative engagement with community and institutional partners is therefore important for making that influence more transparent and for creating opportunities to revise analytic and design decisions throughout the process.

Finally, two key additional process-oriented insights that emerged from the research for both dashboard prototypes are that the cross-sector collaboration essential for dashboard effectiveness requires sustained relationship-building and institutional commitment that extends beyond individual project timelines, and sustainable dashboard infrastructure maintenance requires ongoing technical support, data management capabilities, and stakeholder engagement processes that may challenge resource-constrained organizations.

4.3 Emerging directions: policy implications and future technologies

The dashboards generated several important insights for policy development across multiple domains. Cities and local governments should engage in more targeted outreach by agencies and service providers in neighborhoods under-enrolled in public benefits programs, using the spatial analysis capabilities demonstrated in our dashboards to identify specific areas where gaps between estimated eligibility and current participation rates suggest opportunities for intervention. Understanding travel patterns to food pantries and variation in food costs suggests opportunities for better coordination of emergency food services and more strategic approaches to food assistance that consider both geographic accessibility and service hours. An integrated view of food system components across different public agencies supports more coordinated approaches to addressing food insecurity that consider housing costs, transportation, and other determinants of food access rather than treating food assistance in isolation.

Several opportunities emerged for new technologies to enhance food system monitoring capabilities. While we did not use AI to build the two dashboard prototypes or analyze the underlying data, AI-assisted methods may offer opportunities to support selected aspects of working with large qualitative datasets, such as organizing or identifying potential patterns for further examination. However, particularly in participatory research centered on community voice and lived experience, such tools should complement rather than replace researcher-led interpretation. Qualitative analysis depends on contextual knowledge, reflexivity, and attention to meanings and relationships that automated approaches may not adequately capture, while the use of generative AI also raises concerns related to privacy, bias, transparency, data ownership, and researcher accountability (Davison et al., 2024; Williams, 2024). Any future application of these tools to community-generated data should therefore maintain meaningful human oversight and explicit safeguards for participant data and interpretive integrity.

Participatory data platforms thus represent an area where new technologies could enable more systematic collection of community-generated data and facilitate ongoing engagement in metrics development, moving beyond one-time surveys to continuous community input mechanisms. Real-time data sources, including IoT sensors and other continuously updated information streams, could provide more timely information about food system conditions, enabling more responsive policy adjustments and intervention timing.

While the approaches developed here showed promise in NYC and Long Island settings, replication in other contexts would require careful adaptation to local data environments, institutional contexts, and community priorities. The resource-intensive participatory development process may not be feasible in all municipal contexts, suggesting the need for scalable approaches that maintain community engagement while reducing resource requirements. Opportunities for scaling and replication should focus on developing frameworks that can be adapted to different urban contexts while maintaining the core principles of equity-focused, community-engaged monitoring.

The potential usefulness of participatory food system dashboards also depends on recognizing that they serve multiple audiences with different information needs and capacities for action. Government officials may use disaggregated indicators to identify geographic or demographic disparities, examine program participation and service gaps, and inform policy or program planning. Community-based organizations and service providers may use the same information to understand patterns of unmet need, coordinate services, support advocacy, and strengthen funding or resource-allocation decisions, while funders and local conveners may use dashboard evidence to identify regional priorities and inform grantmaking. For advocates, community organizations, and other public users, dashboards can provide a means of interpreting and questioning official indicators, bringing community knowledge into dialogue with administrative data, and identifying locally salient priorities. Designing for these different audiences therefore requires attention not only to what information is presented, but also to the context, level of detail, and forms of interaction through which different users can meaningfully interpret and apply it.

Accessibility should likewise be treated as a core design consideration in the development of public-facing food system dashboards, particularly when these tools are intended to broaden participation in data interpretation and decision-making. The prototypes incorporated explanatory text, navigation features, and iterative stakeholder feedback intended to improve clarity and usability; however, the projects did not systematically evaluate accessibility against formal standards or assess whether the dashboards were equally usable across different levels of digital or data literacy. Multilingual access was also not systematically incorporated or evaluated, representing an important limitation for tools intended to engage diverse communities, particularly in multilingual urban and regional contexts. Future dashboard development should therefore incorporate accessibility and language access from the outset, including user testing with intended audiences and attention to plain-language communication, multilingual content, and established digital accessibility standards.

Finally, because both dashboards were developed as prototypes, this study does not claim to measure causal effects on enacted policies, budgets, or program outcomes. Instead, we document intermediate forms of policy relevance: the use of dashboard development to structure conversations among government, nonprofit, and community stakeholders; identify priority geographies and program gaps; develop use cases for government officials, service providers, and community advocates; and support local partners’ planning, advocacy, and fundraising. Future research should assess whether and how these intermediate uses translate into policy adoption, resource allocation, and measurable changes in food system equity.

4.4 Actionable recommendations

Based on our findings, we offer several recommendations for policymakers and researchers across multiple domains. Municipal and other local government entities should expand granularity of urban food metrics by investing in neighborhood-level data collection and analysis capabilities that can reveal equity gaps obscured by city-wide averages. They should also integrate lived-experience data into dashboards through systematic community engagement processes that recognize community organizations and residents as essential partners in data interpretation and validation. They should foster multi-sector partnerships for data creation and maintenance by establishing formal collaboration mechanisms between government agencies, community organizations, and research institutions that support ongoing dashboard development and refinement.

Researchers and dashboard developers should develop frameworks for systematic replication of participatory dashboard development processes that can be adapted to different urban contexts while maintaining core principles of equity and community engagement. They should continue to innovate in data collection and analysis methods that center community knowledge, exploring new approaches to integrating quantitative administrative data with qualitative community insights. Researchers and dashboard developers should also incorporate accessibility and language access as explicit design criteria, including plain-language communication, multilingual content where appropriate to the intended audience, formal accessibility standards, and usability testing with diverse users.

For SDG monitoring at the international level, frameworks should be expanded to incorporate narrative and qualitative indicators systematically, providing guidance for how community voices and lived experience data can complement traditional quantitative metrics. International frameworks should also provide guidance guidance for integrating structural determinant analysis into food security monitoring, recognizing that upstream factors like housing costs and transportation access are essential components of food system equity. They should create mechanisms mechanisms for incorporating community-defined metrics into international frameworks, ensuring that global monitoring systems can capture locally relevant indicators that reflect community priorities and experiences.

4.5 Theoretical contributions to data visualization practice

This research contributes to evolving theories of participatory data visualization and community-centered dashboard design. Our findings support the importance of moving beyond user-centered design approaches that focus primarily on usability toward human-centered design that engages broader stakeholder networks and incorporates community knowledge systems (Sanders and Stappers, 2008; Costanza-Chock, 2020; Dörk et al., 2020). The integration of quantitative administrative data with qualitative community insights demonstrates how data dashboards can serve as platforms for knowledge co-production rather than simply information dissemination, consistent with Corburn's (2003, 2005) argument that local knowledge must be integrated into environmental and public health decision-making. Involvement of community stakeholders in the process of dashboard development and the resulting tool can then serve as a means of empowerment and capacity building for local communities (Anderson et al., 2023; Sharp et al., 2022; Stephens and Altamirano, 2022).

The two dashboard projects provided empirical evidence for how different data presentation strategies influence stakeholder interpretation of food system challenges. Including community narratives alongside quantitative indicators reveals cases in which official data patterns do not align with lived experiences, suggesting the need for more sophisticated approaches to data validation that incorporate multiple ways of knowing. The positive response to community quotes integrated within dashboard visualizations supports theories about the importance of narrative framing in data interpretation.

4.6 Limitations and future directions

Several limitations emerged from this work that identify critical areas for future development. While the prototype dashboards establish some of the building blocks for scenario-based analysis, current capabilities remain largely descriptive rather than predictive. Developing robust interactive modeling capabilities would require integration of additional datasets on policy implementation timelines and costs, development of statistical models that can reliably predict intervention impacts, creation of user interfaces or AI-enabled tools that enable non-technical users to modify parameters and interpret results, and validation processes to ensure scenario modeling accuracy and policy relevance.

Certain technical and methodological advances are required to enable dashboards to provide enhanced functionality. Moving beyond static snapshots to incorporate continuously updated data streams would require ample investments in data infrastructure and automated processing capabilities. Machine learning approaches for pattern recognition and predictive modeling would enhance analytical capabilities but require specialized technical expertise and computational resources. Expanding community engagement beyond current stakeholder networks, including through citizen science and large-scale citywide data collection (e.g., for food prices at grocery stores), would require development of more accessible platforms for ongoing data contribution and feedback. The participatory process itself also does not eliminate differences in technical expertise, institutional resources, or agenda-setting capacity among university researchers and other project partners. Future work should examine more explicitly how these differences shape indicator selection, interpretation, and dashboard design.

Further research is also needed to understand how dashboard insights translate into policy changes, funding availability for non-government organizations, city budget allocation decisions, and program improvements. Current evidence suggests that data visualization alone is insufficient to drive policy change without complementary advocacy, institutional support, and political will (Freudenberg et al., 2018; Kitchin et al., 2016). This highlights an important gap in understanding the pathways through which improved monitoring tools may contribute to tangible improvements in food system equity and sustainability.

5 Conclusion

The applied dashboard development research presented in this paper demonstrates the potential for cities to strengthen both monitoring of and progress toward the UN SDGs through locally relevant, equity-oriented food system metrics. The development of the NYC and East End food equity dashboards shows how university researchers, government stakeholders, and community partners can collaborate to create monitoring tools that integrate diverse data sources, center equity, and support evidence-informed decision-making. Our experience also suggests that university partners can provide important convening, analytic, and technical capacity across fragmented data systems, while this intermediary role requires reflexivity about agenda-setting power and sustained accountability to partner priorities.

The two pilot dashboards revealed dimensions of food system equity that may be obscured by conventional monitoring approaches, including spatial mismatches between food need and food assistance, the importance of factors beyond proximity in shaping food access, and opportunities for more targeted interventions. Integrating community knowledge with administrative and novel data sources, while enabling interactive exploration across indicators and geographies, illustrates how local monitoring can move beyond filling individual data gaps toward connecting existing indicators in ways that support more systemic understanding and policy action. Although the prototypes were not originally organized around the SDGs, their retrospective comparison with the SDG framework demonstrate how locally defined indicators can complement global monitoring by adding greater geographic, contextual, and equity-related specificity.

Future research should test and adapt these approaches to additional contexts, develop frameworks for systematic replication, and continue improving methods for integrating community knowledge with quantitative data. The promise of participatory, technology-enhanced food system monitoring lies not only in better measurement, but in creating data infrastructure that supports collective interpretation, accountability, and action toward more equitable food systems. As the 2030 SDGs deadline approaches, strengthening locally-grounded, community-engaged monitoring can contribute to broader efforts to create more just, resilient, and sustainable urban food systems.

Generative AI statement

The author(s) declared that Generative AI was used in the creation of this manuscript. During the preparation of this work, after the first draft was put together, the authors used Claude and ChatGPT in order to improve the readability and language of the manuscript. After using these tools/services, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fsufs.2026.1741924/full#supplementary-material

Compendium of data views from the Food Environment Equity Dashboard – NYC (FEED-NYC), illustrating data on school food, SNAP, emergency food assistance, and food affordability. Supplementary material at the source

Compendium of data views from The East End Food Equity Dashboard, illustrating data on poverty, school food, SNAP, food pantry access, and grocery price variation alongside community survey and interview findings and town profiles. Supplementary material at the source

Footnotes

Edited by: Albie F. Miles, University of Hawaii–West Oahu, United States

Reviewed by: Christopher M. Bacon, Santa Clara University, United States

Ahmad Wahdat, Purdue University, United States

Footnotes

https://a816-dohbesp.nyc.gov/IndicatorPublic/

https://data2go.nyc/

https://experience.arcgis.com/experience/6a3da7b920f248af961554bdf01d668b

Food Environment Equity Dashboard - NYC (beta), https://sites.google.com/view/feed-nyc/home

The East End Food Equity Dashboard, https://sites.google.com/view/aftee-licf-dashboard/home