What You Should Know When Working with Visualization Researchers
An Open Letter to Domain Experts
As visualization researchers, we are fundamentally tool builders. Our field — a subfield of computer science and human-computer interaction — is concerned with creating interactive graphical representations of data to aid understanding, analysis, and decision-making. Working with domain experts is not just part of what we do; it’s central to our mission. Our job is to help facilitate your work as domain experts — be it oncology, civil engineering, journalism, finance, or field biology — through these tools and representations.
This brief letter aims to help domain experts better understand what it means to work with visualization researchers, setting the stage for more productive collaborations. Our partnerships can be incredibly rewarding, but they can also be challenging when expectations don’t align.
Understanding Our Different Worlds
We recognize that many domain scientists operate in high-stakes environments. Your work often involves critical — sometimes life-and-death — situations, and you’re frequently working with limited resources. You barely have enough time in the day to complete your regular job duties, let alone engaging with a group of computer scientists. This reality shapes your priorities and needs, and we deeply respect that. We can help you — if you let us.
As visualization researchers, while our work may not carry the same immediate urgency as yours, we too operate within our own constraints and responsibilities. We are university researchers with obligations to advance scientific knowledge, mentor students, and contribute to our field’s body of research.
What We Are (And Aren’t)
First, it’s important to understand that we are not a software engineering firm. Our primary goal is to generate knowledge, not production-ready code. While we often create open-source work, our focus is on scientific research and innovation in visualization techniques.
We can’t:
- Build production-ready websites or apps;
- Provide enterprise-level software solutions; or
- Guarantee long-term maintenance of tools.
We can:
- Explore novel visualization approaches;
- Conduct research that advances both our fields; and
- Create proof-of-concept implementations.
While our primary output is knowledge rather than code, this knowledge has tangible benefits: it advances understanding of how to effectively visualize data in your domain, which can lead to better tools and approaches in the future. Sometimes, our work does result in code that you can actually use, even if that’s not our primary goal.
The Discovery Process
An essential early part of our collaboration involves what we sometimes call “data therapy” — a deep dive into understanding your data ecosystem. We’ll ask numerous questions about your data: How is it collected? How is it used? Who uses it? What decisions does it inform? What are the critical insights you need to extract? This process helps us understand not just the data itself, but its context and importance in your work.
In fact, many times I wish that domain experts would reach out to visualization researchers earlier in the process. Instead of merely viewing us as the last step to communicate your results — which is undoubtedly one of the strengths of visualization—I would hope you will see us as valuable partners in the analysis process itself. Interactive visualization tools excel at exploratory data analysis, complementing the confirmatory approaches common in most data analysis. By bringing visualization expertise into projects earlier, we can help shape the data exploration journey, potentially uncovering unexpected findings or raising important questions that could influence the research direction.
What to Expect
Much of our collaborative work will involve basic software development, data processing, and framework building that isn’t novel from a visualization research perspective. This “grunt work” is necessary to create a foundation where we can make novel visualization contributions. Please be patient with us if we insist on developing something that may seem trivial or minor in scope; this could well be what we can actually publish as novel visualization research.
Certain aspects of software development — like data transformation, interface polishing, or hardware compatibility — typically don’t advance our research goals. However, we understand their importance and often will do what’s necessary to make the software sufficiently complete and useful. That said, we may not be able to address every issue or account for every possible use case. We have to balance completing the engineering work with maintaining our research focus.
Finding Common Ground
The key to successful collaboration is identifying where our interests overlap. While working with us might not immediately benefit your day-to-day operations, it can offer valuable long-term insights and tools for your field.
Good collaboration requires:
- Patience in developing a common language;
- Grace in handling initial misunderstandings; and
- Understanding that both parties bring valuable expertise to the table.
Your domain expertise is crucial to our work. The more we understand about your field, data, and tasks, the better we can serve both our research goals. However, remember that while we are trained scientists, we likely lack specific expertise in your field. We’ll need your guidance to understand the nuances of your domain.
Moving Forward Together
We absolutely want to help you do better in your field — that’s a core part of our mission. However, successful collaboration requires viewing each other as equal partners. Working together means give and take from both sides. The most successful collaborations happen when both parties maintain clear communication about expectations and limitations, respect each other’s expertise and constraints, and approach the partnership with patience and understanding.
Let’s work together to bridge the gap between visualization research and domain expertise, creating meaningful advances for both our fields.
