Co-design is becoming more common in HCI and visualization, but there is still limited guidance on how to make it work in practice. In this project, we share a two-and-a-half-year co-design collaboration with a local arts community, centered on helping people explore and make sense of how arts funding is distributed.
Through a series of iterative co-design sessions, visualization researchers and community members built shared understanding and developed visualization prototypes tailored to community needs. Along the way, the work rarely felt “finished” – we kept circling back to the messy, uncertain beginnings of the co-design process.
We tell this ongoing story through comic-style visuals and reflect on three recurring “fuzzy front ends” that shaped the project. By sharing these experiences with the visualization community, we offer practical insights for others doing community-engaged co-design.
The co-design began with two visualization researchers (V1 and V2) and two Vancouver Island–based artists (A1 and A2) exploring potential ideas for visualizing arts funding data. Among them, V1, A1, and A2 were the main co-designers during this phase.
V1 also focused on exploring the funding application process: how people apply, what barriers they face, and where challenges emerge. Through this work, mutual understanding gradually developed and a shared vision of the project began to take shape.
V1 also worked on exploring the funding application process: how people apply, what are the barriers that people may face while applying, etc. Through this process, mutual understanding gradually developed, and a shared vision of the project began to emerge.
Once the direction was more clearly defined, V1 began building prototype visualizations, with valuable feedback and suggestions for refinements from A1 and A2 through regular meetings. The process involved a great deal of back-and-forth.
The iterative process helped the team build common ground between visualization researchers and domain experts and navigate the fuzziness of data representation decisions, i.e., using the size of circles to encode arts funding data. Initially unfamiliar with data visualization, A1 and A2 responded with enthusiasm when V1 presented early prototypes using Tableau. Their engagement grew steadily, ultimately contributing to the development of several Tableau dashboards.
Though the dashboards were inspiring, limitations in customization and interactivity became apparent as the co-design deepened. Also, the team identified a more comprehensive and arts-relevant dataset. V3 joined the project to help develop a tailored, interactive visualization for the new dataset.
Through several data understanding sessions, V3 guided the team through the dataset, supporting collective sense-making. Drawing on their experience and expertise, A1 and A2 shared insights into which aspects of the dataset would be meaningful and interesting for the arts sector.
We then brainstormed interaction methods suited to the community’s needs through multiple feedback sessions. The explorations eventually led to the concept of “data painter,” which enabled users to freely “paint” the landscape of Canada's arts funding distribution using a virtual brush.
After the data painter idea was finalized, V3 began developing the tool. The process was iterative, with ongoing feedback and suggestions for low-level features (e.g., zoom, pan, inset maps...) from the rest of the team.
Months of work culminated in a web-based visualization tool that facilitates users to freely explore Canada’s arts funding landscape. V3 showcased the tool to a broader audience, primarily from the arts sector, at a local tech festival to gain more feedback.
The response from the showcase was generally positive, with valuable insights offered for future improvements. V3 shared the feedback with the team. A few meetings followed to discuss possible refinements—until A2 and V3 began to question the value of the dataset itself, marking a crucial turning point in the project.
In the follow-up discussions after the showcase, particularly through a series of focused workshops between A2 and V3, we revisited the data painter prototype to reflect on its scope and limits. This led to the realization of the "dark side" of existing arts funding data, they all excluded unfunded artists and groups.
V3 and A2 shared this important finding with the team. We then went through a series of iterative ideation meetings to explore approaches to surface this missing dimension of existing data. We converged on a two-pronged approach: crowdsourced surveys and expert interviews.
To ensure the survey was well-structured and spoke the right language, V3 held in-person meetings with A2 and A1. These discussions centered on how to elicit and interpret information from artists and arts organizations regarding their funding application experience, especially failed applications.
At the same time, the team became intrigued by the “dark side” concept and its implications for visualization. Extending the idea beyond just data, we held a three-hour brainstorming session to explore and discuss potential dark sides across the entire visualization pipeline.
With support from A1 and A2, V3 completed and published the survey. We reached out to as many stakeholders in the arts sector as possible. Submissions trickled in, some heartfelt, some frustrated, all insightful. Meanwhile, A2 began recruiting and interviewing domain experts.
To date, data collection is still ongoing. We have received 18 survey responses and interviewed a few domain experts. Qualitative analysis has begun, and some themes have emerged, but the final findings remain uncertain. The co-design journey remains fuzzy, iterative, and seemingly never-ending.