Pick a palette family, preview it on your actual chart type, and copy hex, Python, R, or CSS in one click — 10 colorblind-safe palettes from Okabe-Ito to Crameri Vik, plus presets matching Nature, Science, Cell, NEJM, and PNAS.
Distinct hues for unordered groups — treatments, species, cohorts.
Selected palette
8 colors
Best for: unordered groups, treatment comparisons, multi-line charts.
Figure preview
Same data, redrawn with the selected palette
#000000, #E69F00, #56B4E9, #009E73, #F0E442, #0072B2, #D55E00, #CC79A7
Click any palette card to load it into the generator above
The color-blind safe default
Eight discrete colors designed by Okabe and Ito (2002) so every category stays distinguishable for readers with protanopia, deuteranopia, and tritanopia. The most widely recommended categorical palette in scientific publishing.
→ Up to seven groups in bar charts, scatterplots, and multi-line figures.
Source: Okabe & Ito, 2002
Print-safe qualitative
Paul Tol's bright set: seven distinct colors validated for color blindness and tested against PDF print conversion. The categorical palette of choice when Okabe-Ito's yellow is too soft for projector display.
→ Seven-group survey charts, lecture slides, posters that must reproduce in print.
Source: Paul Tol, 2018
The modern default
Tableau's flagship categorical palette and matplotlib's post-2015 default. Not colorblind-safe — adjacent reds and greens collide for deuteranopes — but visually balanced and widely recognized in dashboards.
→ Internal analytics dashboards and stakeholder decks where review accessibility is informal.
Source: Tableau Software (matplotlib 2.0+ default)
Soft categorical fills
ColorBrewer's pastel categorical set: lower saturation, easier on the eye for figures with large filled areas. Not fully colorblind-safe.
→ Stacked area charts, pie charts, choropleth maps with up to eight categories.
Source: ColorBrewer 2.0 (colorbrewer2.org)
High-category categorical
An algorithmically generated palette of maximally distinct colors for cases where you genuinely need 10 or more categories. The 8 hex codes shown here are the first 8 of colorcet's glasbey_bw_minc_20 (avoids near-black and near-white). Not colorblind-safe, but the only honest answer when other palettes collapse.
→ Cluster plots with 10–20 categories, cell type maps, situations where direct labeling is not feasible.
Source: colorcet glasbey_bw_minc_20, first 8
The sequential workhorse
Perceptually uniform purple-to-yellow gradient with monotonically increasing luminance. Designed to stay legible under color vision deficiency and to survive grayscale conversion. Replaces matplotlib jet for serious work.
→ Heatmaps, density plots, intensity maps, sequential continuous data.
Source: matplotlib van der Walt & Smith, 2015
Colorblind-optimized sequential
Mathematically optimized for deuteranomaly and protanomaly. The dark-blue-to-yellow ramp is monotonic in luminance for both standard and most CVD vision, and the palette reproduces predictably in print.
→ Sequential figures where the audience is known to include colorblind readers.
Source: Nuñez, Anderton & Renslow, 2018 (PLOS ONE)
Warm sequential
ColorBrewer's yellow-orange-red ramp for sequential warm data. Not colorblind-safe — readers with protanopia compress the red end — but unmatched for thermal and risk-tier figures.
→ Risk heatmaps, temperature ramps, severity indices where warmth carries semantic meaning.
Source: ColorBrewer 2.0 sequential
The diverging standard
Red-white-blue diverging palette from ColorBrewer. Each side ramps perceptually toward a neutral midpoint, making it ideal for above- and below-zero data. The most-published diverging palette in research figures.
→ Correlation matrices, log-fold-change heatmaps, anomaly maps.
Source: ColorBrewer 2.0 diverging
The scientific diverging
Fabio Crameri's perceptually uniform diverging colormap, designed for geophysics but adopted widely in neuroscience and climate science. Colorblind-safe by construction, with symmetric luminance about the midpoint. The 5 hex stops shown here are evenly sampled from the official 256-color vik.txt distribution.
→ Geophysical anomaly maps, neuroscience activation maps, any diverging plot that must pass CB review.
Source: Crameri Scientific Colour Maps, vik (5-stop sampled)
A scientific color palette is a carefully selected set of colors optimized for research figures. Unlike decorative or brand palettes, scientific palettes are designed against three constraints: accessibility (legible for color-vision-deficient readers), perceptual uniformity (equal numeric steps map to equal visual steps), and reproducibility (consistent across displays, print, and grayscale handouts).
Every palette in this generator has been peer-reviewed in research papers or methods journals. Okabe-Ito (2002), Paul Tol's qualitative sets (2018), Viridis and Cividis (2018), Fabio Crameri's Scientific Colour Maps (2018, 2020) — every one comes with a published rationale and a citation, not a designer's aesthetic preference.
Cross-reference data type and category count to a recommended palette
| Data type | Categories | Recommended palette | CB-safe |
|---|---|---|---|
| Categorical groups (treatments, species) | 2–7 | Okabe-Ito | Yes |
| Categorical groups, print-heavy figure | 2–7 | Paul Tol Bright | Yes |
| Categorical groups, dashboard | 2–10 | Tableau 10 | No |
| Categorical, high category count | 10+ | Glasbey | No |
| Sequential (heatmap, density) | Continuous | Viridis | Yes |
| Sequential, CB-priority audience | Continuous | Cividis | Yes |
| Diverging (above- and below-zero) | Continuous | RdBu | Yes |
| Diverging, geophysics or climate | Continuous | Crameri Vik | Yes |
Approximately 8 percent of men and 0.5 percent of women have some form of color vision deficiency. Many high-impact journals — including Nature, Science, and PNAS — now require or strongly recommend colorblind-accessible figures in their author guidelines.
Informed by ggsci conventions and observation of recent figures
No high-impact journal publishes an official house palette. The conventions below are pieced together from the ggsci R package (a community-maintained set of palettes inspired by major publishers), the Okabe-Ito categorical palette widely accepted in life sciences, and observation of recently published figures. Treat them as starting points, not as official style guides.
Muted qualitative palettes (Okabe-Ito and close relatives) plus Viridis-family sequentials. The author guide explicitly requires colorblind accessibility, and reviewers cite this guideline in revision requests. Preset source: Okabe-Ito 5-color subset.
Slightly warmer palette mix, heavier use of orange-red diverging gradients. Most figures keep to three or four colors per panel; anything beyond that draws pushback in revision. Preset source: ggsci R package pal_aaas().
Slightly desaturated Tableau-adjacent reds and blues (#D62728, #1F77B4). Viridis or RdBu for heatmaps. Distinctive in print without being aggressively saturated. Preset source: Tableau 10 5-color subset.
Conservative: navy, dark red, gray, with occasional teal. Clinical journals reward palette restraint — editors here actively prefer fewer colors per figure. Preset source: ggsci R package pal_nejm().
The most adventurous of the high-impact set — strong adoption of Viridis-family sequentials and Crameri scientific colormaps. Vik and Roma appear in geoscience and neuroscience figures regularly. Preset source: Viridis 5-stop, used here as a stand-in for the PNAS-style sequential workflow.
The pattern across all of these: the higher the impact factor, the fewer colors per figure. If your draft has eight colors in a single panel, collapse it to three or four and use shape, facet, or direct labeling for the rest. Treat the journal presets above as defensible starting points — none of them is a journal-issued style guide.
Paste-ready snippets for the tools that produce most research figures
Wrap the hex list in ListedColormap for heatmaps, or pass it directly to color= for categorical plots. Seaborn's sns.set_palette() accepts the same list. The generator's Python tab ships the full snippet including the import line.
Use scale_color_manual(values = palette) or scale_fill_manual(values = palette) for categorical data. For sequential data, use scale_color_gradientn(colours = palette) with the Viridis or Cividis hex codes.
Format > Shape Fill > More Colors > Custom > Hex (Excel 365), or RGB conversion on older Excel. Save the palette as a custom theme color set in your Office template for reuse.
Drop CSS custom properties into your dashboard root, or pass the hex list as Vega-Lite's range array. D3 accepts the same list as input to d3.scaleOrdinal().
Paste hex codes into the color picker, then save the palette as a Creative Cloud Library, Figma style set, or Affinity swatch group for repeat use across figures.
What journal reviewers and copy editors flag most often
More than seven categorical colors collapses for almost every CVD type and most projector displays. Split the figure into panels or use facets — adding more hues never works.
The single most common revision request. If two adjacent categories use red and green, deuteranopes see them as the same color. Switch to Okabe-Ito or Paul Tol Bright.
Matplotlib jet and other rainbow palettes are not perceptually uniform — visual contrast spikes in green-yellow and flattens in dark blue. Use Viridis or Cividis instead.
Concentration, density, and dose-response data all need sequential palettes. Using Tableau 10 for a five-step dose response makes the numerical order impossible to read at a glance.
Many readers print figures in black and white. Test your figure in grayscale before submitting — Okabe-Ito and Viridis hold up; pastel palettes collapse to indistinguishable greys.
Stop rebuilding the same color scheme across Python, R, and Illustrator. Pick a palette family, preview it on your actual chart type, and copy the paste-ready snippet — every palette in this generator is vetted for colorblind safety, perceptual uniformity, and journal acceptance before it ships.
Choose Categorical for unordered groups and treatments, Sequential for ordered data like heatmaps and density, Diverging for above- and below-midpoint values, or Journal for presets tuned to Nature, Science, Cell, NEJM, and PNAS figure conventions. Each palette card shows a Colorblind-safe tag so you know what is review-proof.
Switch between grouped bar, multi-series line, and heatmap previews so you see exactly how Okabe-Ito reads on a treatment comparison versus how Viridis reads on a 2D matrix. The preview redraws live as you switch palettes — no exporting, no Photoshop overlay.
Grab a hex list for Illustrator and Figma, a matplotlib ListedColormap snippet for Python, a scale_color_manual() block for ggplot2, or CSS custom properties for D3 and dashboards. Every snippet is paste-ready — imports included, order preserved, no helper libraries to chase.
Bench researchers · Data scientists · Lecturers and slide designers
Pick the same Okabe-Ito for a Cell figure, a Tableau-style executive dashboard, and a colorblind-safe lecture slide — without rebuilding the palette in three different tools.
Nature · Science · Cell · eLife · PLOS · PNAS
Figures get sent back for two reasons: too many colors, or red-green collisions. The generator gives you Okabe-Ito for seven groups or fewer, Crameri Vik for diverging heatmaps that pass colorblind review, and Cividis for sequential data that survives grayscale conversion.
matplotlib · seaborn · ggplot2 · Plotly · Vega-Lite
You don't need another color picker — you need a snippet that drops into ListedColormap, scale_color_manual(), or vega-lite range. Copy the exact line you would have typed: imports included, hex order preserved, ready for pyplot.bar() or ggplot() + scale_*.
Lecture slides · Textbook figures · Poster sessions
Your slides go to a 200-seat lecture hall, an online course replay, and a printed poster — all of them have someone in the room with a color vision deficiency. Pick a palette that holds up on projector, retina display, and grayscale handout.
The Scientific Color Palette Generator collects the categorical, sequential, and diverging palettes that reviewers, editors, and replication-friendly data scientists actually use — with paste-ready code for every tool that produces a research figure.
Okabe-Ito · Tol Bright · Tableau 10 · Set2 · Viridis · Cividis · YlOrRd · RdBu · Vik · Glasbey
The palettes reviewers see most often, with hex codes drawn from their original publications. No proprietary AI-generated palettes that get flagged in review.
Eight percent of male reviewers will see your figure differently
Every palette card shows a Colorblind-safe or Not CB-safe tag next to its name. About 8% of men and 0.5% of women have a color vision deficiency — the tag tells you in two seconds whether you are risking that pool.
Imports included. Hex order preserved.
Copy a matplotlib ListedColormap, a ggplot2 scale_color_manual() block, or CSS custom properties — full snippet, ready to drop into your existing notebook or stylesheet.
Informed by ggsci R package + observation of recent figures
No high-impact journal publishes an official house palette. The five presets are pieced together from public conventions: Okabe-Ito subset for Nature, ggsci pal_aaas() for Science, Tableau subset for Cell Press, ggsci pal_nejm() for NEJM, and Viridis for PNAS. Each preset's source is shown on its card.
Equal numeric steps map to equal visual steps
Sequential palettes that ramp luminance evenly. Without uniform luminance, your heatmap exaggerates whatever value happens to fall in the middle of the color scale — the source of most figure misreads.
Cividis is not behind a paywall.
No paywall on any palette. No pro tier for Crameri Vik. No sign-up wall before you can copy the hex. Bookmark the page and use it the next time you open a notebook.
The Scientific Color Palette Generator is free — every palette, every snippet, no sign-up. If you also need AI to draw the figure itself (graphical abstracts, conceptual frameworks, phylogenetic trees, fishbone diagrams), Vizcept Pro covers the rest of the toolkit.
Student or teacher? Apply for an education discount
What researchers, data scientists, and educators say about Vizcept's scientific color palettes.
Dr. Priya Raman
Postdoctoral Fellow, Computational BiologyI used to keep a sticky note of the Okabe-Ito hex codes on my second monitor. Now I copy the ListedColormap line and paste it into the notebook — three seconds, no typos, no off-by-one in the order.
Lukas Weber
PhD Candidate, GeophysicsThe Crameri Vik preset alone saved me a revision round. My anomaly map looked fine on my laptop until a colorblind reviewer flagged the diverging palette. The CB-safe tag pointed me to the right answer in under a minute.
Sarah Chen
Senior Data Scientist, Healthcare AnalyticsOur slide deck has to look the same in Tableau, matplotlib, and PowerPoint. I copy the hex list once, paste it everywhere, and the brand-adjacent palette never drifts between tools.
Dr. Priya Raman
Postdoctoral Fellow, Computational BiologyI used to keep a sticky note of the Okabe-Ito hex codes on my second monitor. Now I copy the ListedColormap line and paste it into the notebook — three seconds, no typos, no off-by-one in the order.
Lukas Weber
PhD Candidate, GeophysicsThe Crameri Vik preset alone saved me a revision round. My anomaly map looked fine on my laptop until a colorblind reviewer flagged the diverging palette. The CB-safe tag pointed me to the right answer in under a minute.
Sarah Chen
Senior Data Scientist, Healthcare AnalyticsOur slide deck has to look the same in Tableau, matplotlib, and PowerPoint. I copy the hex list once, paste it everywhere, and the brand-adjacent palette never drifts between tools.
Dr. Priya Raman
Postdoctoral Fellow, Computational BiologyI used to keep a sticky note of the Okabe-Ito hex codes on my second monitor. Now I copy the ListedColormap line and paste it into the notebook — three seconds, no typos, no off-by-one in the order.
Lukas Weber
PhD Candidate, GeophysicsThe Crameri Vik preset alone saved me a revision round. My anomaly map looked fine on my laptop until a colorblind reviewer flagged the diverging palette. The CB-safe tag pointed me to the right answer in under a minute.
Sarah Chen
Senior Data Scientist, Healthcare AnalyticsOur slide deck has to look the same in Tableau, matplotlib, and PowerPoint. I copy the hex list once, paste it everywhere, and the brand-adjacent palette never drifts between tools.
Dr. Priya Raman
Postdoctoral Fellow, Computational BiologyI used to keep a sticky note of the Okabe-Ito hex codes on my second monitor. Now I copy the ListedColormap line and paste it into the notebook — three seconds, no typos, no off-by-one in the order.
Lukas Weber
PhD Candidate, GeophysicsThe Crameri Vik preset alone saved me a revision round. My anomaly map looked fine on my laptop until a colorblind reviewer flagged the diverging palette. The CB-safe tag pointed me to the right answer in under a minute.
Sarah Chen
Senior Data Scientist, Healthcare AnalyticsOur slide deck has to look the same in Tableau, matplotlib, and PowerPoint. I copy the hex list once, paste it everywhere, and the brand-adjacent palette never drifts between tools.
Everything researchers ask about scientific color palettes
Vizcept's AI generators turn a paragraph of your manuscript into a graphical abstract, conceptual framework, phylogenetic tree, or fishbone diagram — every chart pre-colored with the palettes on this page.
Explore more AI diagram tools across subjects — from the same discipline and beyond.
One subscription unlocks every Vizcept AI diagram tool — from free body diagrams and syntax trees to biology diagrams and more.
