Designing Scientific Figures and Data Visualizations for Presentations
A figure that works perfectly in a published paper can fail completely projected across a conference room. Presentation figures need their own design considerations, distinct from a paper’s static, printed figures, and understanding these differences protects your data from being misread or simply invisible to your audience.
Colorblind accessibility isn’t optional
A meaningful percentage of any audience has some form of color vision deficiency, most commonly red-green. Using colorblind-safe palettes, and never relying on color alone to distinguish data series without a secondary cue like line style or direct labeling, ensures your data is actually readable by everyone in the room.
Simplify beyond what feels comfortable
A figure with six overlapping data series that reads clearly on a monitor at close range often becomes an indecipherable tangle projected across a large room. Reducing the number of series shown at once, or breaking a complex figure into a simpler sequence across multiple slides, usually communicates more effectively than cramming everything into one dense visualization.
Font size needs to account for room size and viewing distance
Axis labels and legends that are perfectly legible on your laptop screen are often unreadable from the back of a large conference room. Testing figures at actual presentation size, or erring toward larger fonts than feel necessary when designing at your desk, avoids this common and entirely avoidable problem.
Statistical significance needs clear, honest visual representation
Error bars, confidence intervals, and significance markers should be genuinely informative, not decorative, misleading or unclear representations of uncertainty in a data visualization can distort how an audience interprets your actual findings, worth double-checking that your visual representation of statistical significance is both accurate and clearly explained.
Consistency across your full presentation aids comprehension
Using the same color coding, axis conventions, and visual style across multiple figures in a single talk helps your audience build a consistent mental model as they follow your argument, rather than having to reorient to new visual conventions with every new slide.
Animation can clarify complex data, used sparingly
Building up a complex figure step by step, revealing one data series or comparison at a time, can help an audience follow a genuinely complex argument that would overwhelm them if shown all at once, worth using specifically for this purpose rather than as decorative motion.
A scientific figure design checklist
- Colorblind-safe palette used, with secondary visual cues beyond color alone
- Complexity reduced or broken into sequential slides for room-scale legibility
- Font sizes tested at actual presentation scale, not just desktop viewing
- Statistical significance represented accurately and clearly explained
- Visual conventions kept consistent across the full presentation
Frequently asked questions
What are reliable colorblind-safe color palette resources?
Several established color palette tools and libraries are specifically designed and tested for colorblind accessibility, worth using rather than manually selecting colors and hoping they’re distinguishable.
How can I test whether my figures will be legible in an actual conference room?
Projecting your slides in a large room before the actual talk, or having a colleague view them from a genuine viewing distance, catches legibility problems a desktop review misses.
Should presentation figures differ from the figures in my published paper?
Often yes, presentation figures generally need to be simpler and larger-scaled than a paper’s more detailed, static figures designed for close reading.