Why this matters
As a BI Analyst, your charts drive decisions. If colors mislead or people cannot see differences, insights are missed. Real tasks include designing KPI cards that read well on projectors, building colorblind-safe categorical legends, ensuring text/line contrast for dark mode, and making dashboards understandable when printed in grayscale.
- Reduce errors: color encodes meaning (e.g., loss vs. gain). Misuse can flip interpretations.
- Reach more users: about 8% of men and 0.5% of women have some color vision deficiency.
- Fewer support questions: accessible palettes and labels make dashboards self-explanatory.
Concept explained simply
Color does two main jobs in data viz: categorize things (A vs. B) and show magnitude (low to high). Accessibility means everyone can decode those jobs, even with color vision differences, on different screens, and in print.
Mental model: Signal, Not Decoration
Think of color as a highlighter pen for your message, not as confetti. First decide the message (trend, change, exception). Then apply the smallest amount of color that makes that message obvious. Always add a backup cue (text, position, size, pattern) so meaning survives when color fails.
Core rules you can apply today
- Use color sparingly for emphasis; keep most elements neutral gray.
- Never rely on color alone. Add labels, icons, markers, patterns, or position.
- Pick the right palette type:
- Categorical: distinct, equally salient colors (up to ~6–8 categories).
- Sequential: one hue, light to dark for ordered values.
- Diverging: two hues splitting around a meaningful midpoint (e.g., 0).
- Avoid rainbow palettes for ordered data. They create false jumps.
- Contrast: aim for at least 4.5:1 for normal text and thin lines; 3:1 for large text/bold lines.
- Test in grayscale; your chart should still make sense.
- For alerts, pair color with an icon and text (e.g., “Error”).
- Be consistent across charts (same category, same color, everywhere).
Quick grayscale check (no tools)
- Squint at the chart or view on a low-brightness screen.
- Ask: can I still rank values, find the highlight, and read text?
- If not, increase contrast, add labels, or reduce saturation.
Worked examples
Example 1 — Fix a rainbow heatmap
Problem: A rainbow heatmap of churn probability makes mid-range values look like distinct groups.
Fix: Use a single-hue sequential palette (e.g., light to dark blue). Add value labels for the top 5 hot spots. Keep gridlines subtle gray. Result: viewers perceive a smooth progression; the highest risk pops clearly.
Example 2 — Multi-line chart for color-blind users
Problem: 5 product lines share red/green/blue shades; two lines are indistinguishable for deuteranopia.
Fix: Assign colorblind-safe distinct colors. Add redundancy: different dash patterns and end-of-line labels. Increase line thickness slightly. Result: any viewer can track each line, and labels reduce legend lookups.
Example 3 — KPI card contrast
Problem: Gray text on a pastel background fails contrast on laptop projectors.
Fix: Use dark text (#1A1A1A) on white (#FFFFFF) or near-white. For dark mode, use near-white text (#EDEDED) on very dark gray (#121212). Keep accent color for small highlights, not the text body. Result: Readable KPIs in all rooms.
How to choose palettes fast
- Decide the message type: category, order, or split around a center.
- Start neutral: gray for most elements; add 1–2 accent colors for highlights.
- Limit categories: if >6–8, group or switch to labels/patterns/position.
- Check contrast once for text/lines; adjust until 4.5:1 or better.
- Do the grayscale test; add labels if meaning fades.
Accessibility checklist
- Text and thin lines meet ~4.5:1 contrast against background.
- Large text (≥18pt or bold ≥14pt) meets ~3:1 contrast.
- Color is not the only cue (use labels, icons, patterns, or position).
- Avoid red/green conflicts; choose colorblind-safe sets.
- Same category = same color across the dashboard.
- Directly label lines/bars where possible; minimize legend hunting.
- Interactive states (selected/hover/focus) differ by more than color (e.g., underline, thicker border).
- Charts remain understandable when printed in grayscale.
Exercises
Try these in your BI tool or on paper. Hints and solutions are provided below each task.
Exercise 1 — Build a colorblind-safe categorical palette (ex1)
Design a palette for 6 product categories and apply redundancy so the chart works for color-blind users and in grayscale.
- Deliverables: 6 hex colors, where/how to label, and any non-color cue you’ll add.
Hints
- Prefer medium saturation colors that differ in hue and lightness.
- Add direct labels on bars/lines to reduce legend dependence.
- Patterns or markers can add redundancy if needed.
Exercise 2 — Contrast-proof a KPI card (ex2)
Create a KPI tile for “Net Profit: 2.8M (+6.2%)” that remains readable on bright projectors and in dark mode.
- Deliverables: background, text, and accent hex values; brief contrast rationale.
Hints
- Use near-black text on white, or near-white text on very dark gray.
- Accent color should not be the body text color.
- Thin lines and small text need higher contrast.
Common mistakes and how to self-check
- Mistake: Rainbow palette for ordered data. Fix: Use sequential or diverging palettes.
- Mistake: Red/green to denote states. Fix: Pair with icons/text; pick colorblind-safe hues.
- Mistake: Too many category colors. Fix: Cap at ~6–8; group or label directly.
- Mistake: Low contrast on text/lines. Fix: Darken text or lighten background until readable; aim for ~4.5:1.
- Mistake: Legends only. Fix: Direct labels or annotations near data.
Self-check mini test
- Convert your chart to grayscale: can you still see the story?
- Hide the legend: can you still identify categories?
- View at 50% zoom: is critical text still readable?
Practical projects
- Redesign a company dashboard: replace rainbow/brand-saturated visuals with neutral base + limited accents; document your palette decisions.
- Accessibility audit: pick 3 existing charts and create a before/after with contrast notes and colorblind-safe improvements.
- Alert system: design a set of status banners (error, warning, success, info) with color + icon + text redundancies.
Who this is for
- BI Analysts and data professionals who design dashboards and reports.
- Anyone responsible for readable KPIs across devices and audiences.
Prerequisites
- Basic chart literacy (bars, lines, heatmaps).
- Ability to adjust colors and labels in your BI tool.
Learning path
- Color and accessibility basics (this lesson).
- Chart selection for the message (trend vs. comparison vs. distribution).
- Annotations and labeling that reduce legend dependence.
- Layout and Gestalt principles to guide attention.
- Interaction states and dashboard consistency.
Next steps
- Apply these rules to one existing dashboard.
- Share before/after examples with a short note explaining each change.
- Take the quick test below to check your understanding.
Mini challenge
Take any chart you’ve made. Remove all color. Add labels, markers, or patterns until it still tells the same story. Now reintroduce just one accent color to highlight the single most important insight.
Quick test
Available to everyone. Sign in to save your progress.