Scientific Visuals

The first theme of the workshop was distinguishing between different kinds of scientific visuals and understanding the conventions that shape them. We examined information graphics, explanatory diagrams, direct data visualizations, and abstract illustrations, each serving a distinct function in science communication. The guiding principle was that “every visualization is a hypothesis”: the way we choose to represent data or concepts already frames how they will be interpreted. This perspective turned even simple decisions about format or style into acts of reasoning rather than matters of aesthetics.

The sessions also highlighted the lineage of scientific visualization. Ernst Haeckel’s intricate drawings of marine life, Dave Goodsell’s watercolor renderings of molecular environments, and Janet Iwasa’s dynamic animations were presented as milestones in the craft. These examples demonstrated how clarity, creativity, and accuracy can combine to make science both accessible and memorable. They also showed that great visuals often outlive the texts they accompany, embedding themselves in the collective imagination of science.

We discussed how to balance abstraction and accuracy, deciding when a schematic is more useful than a literal depiction. Visual metaphors and analogies were introduced as powerful tools, especially when dealing with abstract processes that cannot be photographed or observed directly. The challenge is not simply to illustrate but to make choices that communicate more clearly than a lenghty explanation could.

Design Principles

A major focus of the workshop was the language of design: the rules and habits that make visuals clear, legible, and effective. We began with typography, distinguishing between serif and sans-serif fonts, and learning how choices in line spacing, letter spacing, and kerning influence readability. Subtle decisions such as font thickness, contrast, and color tone were shown to have a strong impact on how an audience perceives text, especially when projected on a slide or reduced in size for a figure.

Color theory was introduced as a foundation for building clarity in figures. We explored how hue, saturation, and lightness can be combined to create palettes, and why limiting the number of colors often strengthens a figure. Attention was given to accessibility: colorblind-friendly palettes such as Viridis or blue-orange schemes provide clarity without excluding part of the audience. We also practiced checking our designs in grayscale, a simple test that often reveals whether contrast is strong enough to stand on its own.

Layout was framed as the invisible structure of a figure. Negative space, hierarchy, and alignment are not decorative but functional, guiding the eye toward what matters. The principle of “amplify the signal, minimize the noise” was repeated throughout, reminding us that clutter is the enemy of comprehension. Even the placement of captions and legends was treated as a design choice, with the potential to either clarify or distract.

Finally, we addressed common myths about illustration in science. Many of us assume that artistic talent, expensive software, or endless time are prerequisites for producing effective visuals. The workshop dismantled these assumptions by showing that clarity depends less on drawing skills and more on decisions about structure and emphasis. With free tools, simple design principles, and iterative feedback, anyone can improve their figures. The message was that high standards in visual communication are within reach, provided we treat design as a deliberate part of the scientific process.

Data Visualisation

Another part of the workshop dealt with the choices we make when presenting quantitative results. We reviewed the strengths and limitations of standard plot types and how each one carries implicit assumptions about the data. Bar charts, for instance, are familiar and easy to read, but they can oversimplify variation and hide distribution. Boxplots offer a clearer view of spread, outliers, and central tendency, while violin plots add the dimension of density, showing the probability distribution at a glance. Line charts excel at tracking trends across time, scatter plots reveal relationships between variables, and heatmaps capture patterns across large datasets. The lesson was that no single format is universally best, but each serves a distinct role depending on the question being asked.

Common mistakes were highlighted as well. Overly cluttered graphs, inconsistent axis scales, default software styles, unnecessary 3D effects, and text that is too small to read all undermine clarity. Many of these errors arise not from a lack of statistical knowledge but from neglecting design principles. Even subtle details, such as the choice of line thickness or the use of too many colors, can distort interpretation or overwhelm the viewer.

The workshop emphasized that figures should not just display numbers but tell a story about the data. Variability should not be hidden behind averages, and scales should remain consistent across related plots to allow fair comparisons. The goal is to guide the audience toward the key point without distracting them with noise. This mindset encouraged us to treat data visualization as both a scientific and a rhetorical act, one where design choices carry as much weight as statistical ones.

Storytelling and Presentations

The workshop also addressed the role of narrative in scientific talks. We often prepare presentations by simply stacking slides full of results, but this approach risks losing the audience in a stream of information. Instead, we were introduced to structures that treat a talk as a story, with a clear rhythm of setup, conflict, and resolution. The And–But–Therefore framework, popularized by Randy Olson, provided a simple template: begin with shared ground, introduce a problem or contradiction, and then move to the solution. This approach creates a natural flow that keeps attention focused on why the work matters.

To illustrate this, the workshop drew on well-known examples such as novels, movies, and even pop songs, showing that the same logic of tension and release applies across genres. The structure of the hero’s journey was also discussed as a way to think about how audiences experience a talk: there is an opening world of background knowledge, a call to a problem, obstacles in the form of unanswered questions, and ultimately the “return” of the solution. While this framework is not applied literally in science, it reminded us that listeners expect some form of transformation by the end.

Practical strategies followed from these ideas. Each slide was framed as its own small narrative, with the conclusion of one slide setting up the next. We discussed the value of rehearsing extensively, trimming information that does not serve the story, and using design to highlight the most important points. The time investment required for a polished talk was made explicit: dozens of hours for figure preparation and slide design, followed by repeated rounds of rehearsal.

The underlying message was that effective presentations are not about being flashy but about aligning content, design, and delivery into a coherent narrative. A well-structured talk does not just communicate results, it makes the audience care about them. Storytelling, far from being a distraction, was presented as a scientific tool in its own right, one that helps ideas resonate and endure.

Creative Extensions

Toward the end of the workshop we turned to more experimental approaches to scientific visuals, looking at ways creativity can expand the reach and impact of research. One topic was watercolor-style figures, where hand-drawn outlines are combined with colored washes and gradients to create images that feel both precise and organic. These techniques are not about abandoning accuracy but about presenting scientific ideas in a style that draws in the viewer and invites closer attention.

Another theme was the use of generative AI. Tools such as Midjourney were shown as ways to quickly prototype visual metaphors or explore stylistic options. The key was to think abstractly, focusing on shapes, contrasts, or textures that capture the essence of a concept rather than trying to generate literal images of scientific objects. Iterative prompting, testing different art styles, and post-processing were framed as part of a creative workflow rather than a shortcut.

Cover artwork provided a third example of creative extension. We examined journal covers that distilled the message of a paper into a single striking metaphor. These covers increase visibility, shareability, and memorability, often becoming the most recognizable part of a study. The process requires condensing a project into a singular punchline, finding an analogy that resonates, and matching the design style to the expectations of the target journal.

The discussion closed by returning to the broader question of science as a creative act. Scientific models and figures are not reality itself but representations, crafted with choices about what to emphasize and what to leave out. By embracing this fact, we were encouraged to see the overlap between science and art not as a weakness but as a strength. Creative approaches, whether through hand-drawn styles, digital tools, or visual metaphors, can open new pathways for making science understandable and memorable.