AGENDA

Optional networking / eagagement, feedback clinic, etc.

AI-powered coding tools have made it possible to build charts, dashboards, and interactive stories faster than ever. But while vibe-coding accelerates the journey from idea to prototype, it can also create the illusion that the work is finished.

Drawing on the making of Bollywood Falls out of Love with Kontinentalist, this session explores the gap between rapid AI-assisted exploration and a publishable data story. Through a behind-the-scenes look at the project, I examine where AI excelled—from data gathering and analysis to visual prototyping—and where human judgment remained essential: shaping narratives, making design decisions, providing context, and ensuring trustworthiness.

The session asks a simple but urgent question: as AI takes on more of the drafting, what responsibilities still belong to the maker? The key takeaway is that while vibe-coding can be a powerful engine for first drafts, the final draft—and the accountability that comes with it—still belongs to you.

How do we visualize what is not there — data that is incomplete, erased, or deliberately hidden by authoritarian regimes? This talk explores the paradox of representing absence through data visualization, drawing on the experience of the Argentine Forensic Anthropology Team (EAAF) and its project La Búsqueda, alongside other visual experiments around enforced disappearance during Argentina’s last dictatorship.

Unlike conventional datasets, absence itself becomes central: missing identities, destroyed records, interrupted stories, and silenced families. Rather than presenting definitive methods or solutions, the talk proposes a set of reflections and open questions about how to visualize uncertainty, incompleteness, and state-produced voids without erasing their complexity. What does it mean to show data that is not there? And how can visualization make absence itself perceptible?

The Future of DataViz: How AI, D3, and React Are Taking Over

AI is changing the data visualization landscape, making powerful tools more accessible than ever. In this session, we'll explore the major families of dataviz tools, how AI is reshaping them, and why more practitioners are moving toward D3 + React for building advanced, interactive visualizations. 

The talk includes a short live demo and concludes with audience Q&A.

Relax and reset

This talk takes the audience on a journey through the material history of data, from Mesopotamian clay tokens and Inca khipus to analog measuring devices, databooks, and IBM punched cards. Across cultures and eras, humans have developed physical systems to encode, record, and transmit knowledge long before the computer. By bringing together artefacts traditionally studied separately, the talk explores a broader history of material knowledge systems and offers a different perspective on data physicalization today.

How can analysts, researchers, and developers apply an equity-conscious lens to their data analysis and data visualization work? As data increasingly shapes public understanding and policy decisions, the need for inclusive and equitable data practices is more urgent than ever. Just as we carefully consider our written words, we must be equally intentional about how we visually present data to our readers, users, and audiences—including the language, framing, and design choices that surround those visuals.

DVS Program Future

We often talk about data visualizations as if they have a final draft, a moment when the design is complete and the work is done. But that framing breaks down for a specific category of data visualization tools: interactive data explorers. A data explorer is not a one-time narrative visualization. It is an interactive system designed for ongoing use, recurring data updates, and multidimensional exploration. These tools are usually embedded in institutional workflows and depend on evolving datasets, shifting requirements, CMS-driven content models, and performance constraints. In this context, there is no true final draft. Not because teams are indecisive, but because the surrounding system continues to evolve. This talk reframes iteration as a structural necessary property of data explorers rather than a process failure. For designers and engineers, this requires a different mindset. Instead of polishing toward stability, we must architect for change. I’ll share concrete strategies drawn from real-world projects, including: Treating metadata as a design and engineering contract Building configurable chart systems instead of one-off custom components Designing data and architectures that anticipate growth and revision The goal is not to perfect a final artifact, but to build visualization systems that can evolve responsibly over time.

To tell the story of 25 years of offshore wind in the UK, an analyst and a data visualisation designer set out to build something ambitious - a rare chance to craft a rich visual narrative around a climate good news story. Soon, the ideas multiplied. We explored animations, scrollytelling, embedding sounds of the sea, coined phrases like the “Windustrial Revolution,” borrowed heavy metal metaphors, and even ended up in a local maritime museum. In one of these moments of inspiration, 25 charts for 25 years were suggested. The brief started to sprawl — but was it ever pinned down at the start? What ultimately shaped the final outcome wasn’t the hundreds of ideas shared, or a lack of imagination, but partner needs, audience expectations and a very hard external deadline. Those constraints brought us back to the format of a report — but one with distinctive, hand-drawn graphics and a creative flair, built through team effort. This talk reflects on the specificities of think tank analyst–data viz co-creation, creative detours, and how constraints don’t diminish ambition — they refine it.

Relax and reset

Optional networking / eagagement, feedback clinic, etc.

Atomic design isn’t just for UI components, it turns out it’s also a powerful framework for organizing data visualization decisions. This talk goes deep into how a small data viz team used atomic design principles to create final handoff documentation for a client. This documentation was the culmination of wide charting explorations across a large survey research platform’s data. Stress-testing scalability of variable types, complexity levels, and label scenarios ultimately defined the chart parts (atoms) and combinations of these parts (molecules) that became the organizing structure for our Figma documentation system to then be leveraged by both the client’s team of designers and developers.

Electronic music is the origin of my visualization career. Audio production provided me with a background in both the technical details of data, and a foundation for my design process. On the technical side I learned the particulars of data: bit depth, sample rate, signal to noise, and distortion. On the non-technical side I learned how to build a cohesive whole from individual pieces, how to layer elements, and perhaps most importantly, how to iterate and improve on an initial design.

Relax and reset

Data Portraits - Craft Circle

An exploratory session on personal and expressive data visualization. What does it mean to visualize yourself, your community, or your experience? Bring curiosity and a willingness to experiment in whatever tools or materials you have on hand.

Most people don’t have the luxury of spending days polishing a dashboard or turning a dataset into a museum‑ready masterpiece. In real workplaces, analysts, communicators, and program staff are asked to produce clear, decision‑ready visuals in hours—not weeks. This session introduces a practical, accessibility‑first approach to “good‑enough” dataviz: fast, thoughtful, high‑impact visuals built in the everyday tools people already use.  Through real examples of scorecards, one‑pagers, and quick dashboards, you’ll learn how to move from messy data to a clean, comprehensible visual in a single workday. We’ll focus on the essential 20% of design decisions that create 80% of the clarity: choosing the right chart under pressure, applying small edits that dramatically improve comprehension, and prioritizing accessibility when time is tight. You’ll leave with a repeatable workflow, a set of practical shortcuts, and the confidence to deliver visuals that are clear, inclusive, and effective—even when your deadline is “by this afternoon.”

For my MA Final Major Project, I designed #didntreadthetandcs — a one-day interactive exhibition about what happens when people click "I agree" without reading privacy policies. Hand-crafted, participatory, and designed to make visitors active participants in the data.  One design choice led to something I didn't expect.  Part of the experience was a photography waiver with a gotcha clause hidden in it: agree to give me a high five. I tracked how many people actually read it.  At an earlier show, I did a big reveal moment. It landed with laughs, surprise, high fives.  For the final exhibition, I put the gotcha on the information plaque instead. No big reveal, just there for people to find. Far fewer high fives. People who spotted it early didn't want to "ruin" the data.  This talk explores how small decisions about what you reveal (and when) shape how audiences engage with participatory data. Ethical design choices, unexpected outcomes, and lessons I'm carrying into whatever comes next.

This talk showcases the final draft of a data physicalization project - a data quilt about fast fashion made out of repurposed fabrics - and the process on how we made it. Elsie Lee-Robbins and Alexandria Rayburn embarked on what turned out to be a multi-year, cross country crafting data project. The first draft of the project involved choices about what data to use, then drafting the quilt design, and iterating on the choice of fabrics throughout the process. In the end, we created a data visualization quilt that embodies the very context of the data. This quilt could be a data communication tool to start a conversation about how to reduce our climate impact in fashion choices. In our talk, we explore craftivism, unique physical mediums, and visualizing sustainability data for social good.

Relax and reset

Optional mix of structured and open conversations running alongside the main program. Some sessions will have a set topic and host. Others will be shaped by whoever shows up.

In the world of economics and public policy, the PDF is the undisputed king of deliverables. Yet, it is often the place where research and complex data visualizations go to die—static, unchangeable, and frequently misaligned with the needs of those who must use them. This talk argues that by applying Design Methodologies—such as Design Thinking, Sprints, and UX—the public sector can move beyond the "dead" document toward living information systems. 
While these methodologies have revolutionized the private sector, the public sector operates under a unique set of constraints: political optics, rigid funding cycles, and the pressure of year-over-year continuity. We present a practical framework for adapting these strategies to governance. Attendees will learn how to view the PDF not as an end point, but as a drawing board from which dynamic products can emerge, and how to establish a design-led architecture that allows policy products to be revisited and redesigned as user needs evolve. 

When we set out to build Hex's first visualization subagent, we thought the hard part would be the AI. It wasn't. The hard part was visualization.

A good visualization is rarely made in one shot. You generate something, assess whether it's working, and revise. That loop, intuitive to any visualization practitioner, turned out to be exactly what we had to build into the agent. Doing so forced us to answer questions we'd long taken for granted: What does it mean to critically see a chart? How do you get from a first draft to a better one? How do you measure quality when the outcome is inherently non-deterministic and subjective?

Drawing on concrete examples from building a visualization agent that operates inside a real-world AI analytics platform used by thousands of data teams, this talk walks through how we approached each of those questions - and what the process revealed about visualization itself.

Relax and reset

Structured and organic networking sessions and community spaces designed to spark meaningful peer-to-peer connections across skill levels and geographies.

Feedback & Career Clinic (Pre-Booked)

Bring a piece of work for honest, specific feedback. Option to apply as a volunteer reviewer or program support.

Structured and organic networking sessions and community spaces designed to spark meaningful peer-to-peer connections across skill levels and geographies.

Feedback & Career Clinic (Pre-Booked)

Bring a piece of work for honest, specific feedback. Option to apply as a volunteer reviewer or program support.

This talk presents Whale Traces, a research-creation project that explores emotional and accessible data physicalization through individual whale migration data. By combining data cleaning, cartographic experimentation, and physical, multisensory data representations, the project challenges dominant, vision-centered and aggregate approaches in scientific communication. The talk reflects on how design choices can foreground care, storytelling, ecological empathy, and more inclusive ways of accessing data beyond sight alone.

Hear about the obstacles faced and key learnings gained by a visually impaired data visualizer, with practical strategies for accessibility and success.

In 2020 Andy Kirk published his most substantial visualisation work to date: a passion project based on a deep exploration and analysis of every episode of the popular US sitcom 'Seinfeld'. Driven by a curiosity about how characters, laughter and locations were used in the show, he explored how situation comedy can be represented through data and visuals. This original work output was released in the form of 176 copies of a book packed full of analysis (equal to the number of episodes). These limited-edition books were offered to interested people on a pay-what-you-can-afford basis, with all proceeds going to charity. Five years later, Andy began a collaboration with Anne-Marie to reimagine this project for a wider audience as a bespoke digital design solution. No longer bound by the constraints of print, how could the same analysis be conveyed in a fresh way? With a central theme being about the musicality of comedy, the collaboration was extended to Loud Numbers (Duncan Geere and Miriam Quick) to further iterate the possibilities of this data exploration through representing data using sonification techniques. This short talk will offer a whirlwind tour of the key layers of this reimagining process: what we did, why we did it and how we arrived at our final draft.

Relax and reset

What does a “final draft” look like when it must work for the whole world? In the UN Intergovernmental Panel on Climate Change (IPCC), it means visuals shaped through years of iteration, collaboration with 200+ climate scientists, negotiation from 195 governments, and tens of thousands of individual comments. This talk shares the behind-the-scenes of climate graphics that reached the front pages of global media, revealing the processes used to design in highly collaborative environments. Creating these data visuals can feel like designing for hundreds of clients at once, where evolving creative exploration meets the precision of scientific language and the procedural rigor of UN consensus. Ultimately, the ‘final draft’ is a collective achievement: global UN agreement. As creative processes are increasingly incorporated into international and scientific contexts, this session outlines practical tips for navigating complex, high-stakes collaborations and for turning vast amounts of input into clarity, consensus, and -hopefully- climate impact.

Structured and organic networking sessions and community spaces designed to spark meaningful peer-to-peer connections across skill levels and geographies.

Feedback & Career Clinic (Pre-Booked)

Bring a piece of work for honest, specific feedback. Option to apply as a volunteer reviewer or program support.

Optional networking / engagement, feedback clinic, Speaker BTS, etc.

Relax and reset

I trace the evolution of my thinking and motivations as both a data visualization and urban planning practitioner investigating how people access cities for joy and leisure through data visualization, research, travel demand modelling and participatory methods. I present three examples from my practice and research: a web-based data story showcasing how women in South and Southeast Asia access cities for joy and leisure; a multilingual zine and educational resource around how women in India relate to surveillance in public space; lastly, a participatory data sculpture on the theme of nighttime leisure travel, supported by statistical analysis of travel patterns in the Greater Boston area. Each project reflects an iteration of my values, motivations and ethics on the access to leisure in cities, shaped by new geographies, collaborators, and constraints. 

The first project, A Woman’s World: Creating spaces for joy, leisure, and resistance in South and Southeast Asia, was a web-based data journalism project I led while at Kontinentalist (and in collaboration with Khabar Lahariya) that documented how feminist and queer grassroots groups across 10 cities are reclaiming public space through leisure and community-building. Produced as a multimedia data story and social media campaign, the project combined interviews, surveys, and visual storytelling to shift the dominant narrative of victimhood towards collective agency. The second project, A Woman’s Place Is in a Safe City, takes the form of a trilingual zine and educational website developed with grassroots organizations in two Indian cities, in collaboration with STS and Critical Data Studies researchers during my graduate degree. Through participatory action research, we examined the role of surveillance in shaping “safe” public space in India, translating community insights and needs into six actionable areas for public investment beyond policing and monitoring. The third project, When Girls Just Want to Have Fun, How Do They Go?, brings these questions to Boston through a hybrid statistical and artistic approach. Using Massachusetts travel survey data alongside a participatory corkboard mapping exercise, I investigate nighttime leisure mobility to understand motivations beyond safety narratives for travel after dark. 

Each project is framed as a provisional synthesis shaped by curiosity, feedback, and real-world limits. Community interviews reshaped research questions; linguistic and institutional contexts transformed outputs into accessible formats; and material constraints redirected visualization into tactile participation. Progress appears through translation—across cities, disciplines, and media—rather than completion. Each iteration responds to new tools and environments, demonstrating how constraint can clarify intention and deepen inquiry. In this way, the “final draft” is not a perfected end state, but a moment where accumulated learning, critique, and context align—before my next revision inevitably begins.

In an era of misinformation, the integrity of civic discourse depends heavily on the accessibility and perceived neutrality of government data. While data accuracy is foundational, the visual presentation of this information significantly shapes citizen engagement and trust; yet, accessible, un-politicized, and reliable sources remain rare. USAFacts aims to address this deficit as a non-partisan, not-for-profit organization dedicated to making US government data understandable. It was recognized by Ads Fontes Media (2025) as one of the least politically biased and most reliable media sources. However, a critical challenge in this mission is navigating the competing goals of diverse stakeholders—brand, marketing, and product teams—while producing visualizations at scale that remain strictly objective. This talk presents the development of a scalable data visualization design system for USAFacts, created in collaboration with Artefact, a research, strategy, and design consultancy, and launched in the fall of 2025. It demonstrates that by codifying design principles centered on neutrality, recognizability, and accessibility, organizations can empower teams of varying backgrounds to consistently create trustworthy visualizations at scale. The resulting system encompasses all visualization layers—from accessible color palettes to typographic hierarchy—ensuring clarity across media channels ranging from static charts to complex interactive visualizations.

Hang out, relax, and connect as we close out the program

The conference is done. Now it is time to hang out with your fellow Outliers, unwind, and enjoy some light activities together before we close out the conference.

Join us a few minutes early to get set up for success! This informal session is your opportunity to test your audio and video, troubleshoot any technical issues with our support team, and get comfortable in the virtual environment before the program begins. It’s also a great chance to connect with the team, ask any questions, and settle in so you’re ready to hit the ground running when we kick off.

Early arrival is encouraged but not required.

The day kicks off with a brief welcome from the Outlier team. Get oriented on the structure of the day, how the Outlier Data Viz Challenge works, and what to expect as you move through sessions. Today includes two parallel paths for the data and tooling work: Focused, for participants working with a scoped dataset using low to no-code approaches, and Extended, for participants ready to work with a broader, more complex dataset using code-based methods. Both paths produce real, valid work. You will have the opportunity to choose your path during the AI Data Exploration session. Everything before and after that is the same experience for everyone.

Every visualization starts with a real problem, real data, and real constraints. In this session, Heather Page from the United Nations introduces the challenge that will drive the entire day. She will walk through the Sustainable Development Goals (SDG) datasets participants will be working with, explain the communication objectives behind the UN’s 2026 SDG Report, and outline the statistical, methodological, and audience constraints that shape what an effective data viz looks like. Heather will also introduce how the datasets are segmented based on the SDG complexity. 

The SDGs represent a globally agreed framework for measuring progress on the world’s most pressing development challenges, from poverty and health to climate and education. Understanding how that data is collected, what it can and cannot show, and who needs to understand it is where every good visualization decision begins. Come with questions and leave ready to build.

With the challenge brief in hand, this session bridges the gap between understanding a problem and committing to a direction. Jon Schwabish walks through how to move from a complex brief to a concrete visualization concept, covering how to identify the core message, make early structural decisions about purpose, scope, form, format, and use rapid wireframing to test ideas before investing in execution. A key part of this session is understanding your audience: who they are and how much complexity they can absorb shapes every design decision that follows. Participants will leave with a working wireframe, a clear sense of your audience, and a focused set of data questions ready to drive your exploration work.

Understanding your data is one of the most time-consuming parts of any visualization project, and also one of the most consequential. Ray McLendon opens this session with a unified overview of AI-assisted workflows for data exploration, covering the full landscape of available approaches so every participant understands their options before choosing a path. We will start with prompting LLMs for beginners and progress to coding approaches for more advanced analysis. Everyone will walk away with prompt and code examples to use during the workshop and beyond.

The session splits into a structured practical where participants self-select into the track that best fits their experience and goals for the day.

Focused: For participants who want to work with a scoped slice of the SDG data using low and no-code AI tools. Build confidence in the dataset, surface key patterns, and leave with clear insights ready to carry into your design work.

Extended: For participants comfortable with programming who want to work with broader, more complex dataset(s). Go deeper into the data, automate exploration steps, and build a richer analytical foundation to carry into your design work.

A visualization supporting or accompanying a UN report needs to do more than present data accurately. It needs to communicate with clarity, credibility, and visual intention. In this workshop, Martina Zunica covers the core principles of effective visual design, including hierarchy, typography, color, and layout, through exercises applied directly to participants’ works in progress. 

Regardless of which data path participants have been working on, the design principles and feedback in this session apply equally to every draft in the room. The goal is not a finished piece but a meaningfully stronger draft: one that reflects deliberate design decisions and is ready for the feedback and iteration ahead.

At the end of a full day of building, this session brings the room back together to reflect on what emerged. A structured conversation around the approaches participants took across both paths, the decisions that shaped their work, and the tradeoffs that came up across different SDGs. The goal is not to evaluate submissions but to surface shared insights, open questions, and productive tensions that participants can carry into the project review sessions on Outlier Days 1 and 2.

Participants will receive instructions for updating their submission in the portal overnight before reviews begin. Leave the day with momentum and a clear sense of what comes next.