Why this matters
As a BI Analyst, your charts drive decisions. The same data, shown with a poor chart, can hide patterns or mislead stakeholders. Choosing the right chart helps you:
- Spot trends early (e.g., seasonal drops in conversion).
- Compare performance fairly (e.g., product vs. product).
- Explain distributions and risk (e.g., delivery time variability).
- Reveal relationships (e.g., spend vs. revenue).
Real tasks you will face:
- Build an executive KPI dashboard (trend, variance, part-to-whole).
- Analyze an A/B test outcome (distribution and confidence display).
- Explain drivers of churn (ranking + relationships).
Concept explained simply
Charts answer questions. Start from the question, then map to a chart family.
Mental model
Question → Data Type → Chart Family → Design Tweaks
- Compare categories at one point → Categorical values → Bar (horizontal if long labels)
- Change over time → Time series → Line (area for cumulative or uncertainty bands)
- Distribution of values → Numeric set → Histogram, Box, Violin
- Relationship between variables → Two+ numeric → Scatter (add trendline)
- Part-to-whole at one point → Shares → 100% Stacked Bar, Treemap (few vs. many parts)
- Ranking/ordered comparison → Sorted bar/column
- Heat or intensity across two dimensions → Heatmap
- Flow from sources to targets → Sankey (use sparingly, only for flows)
- Geographic patterns → Choropleth or symbol map
Chart families cheat sheet (open)
- Trend: Line > Area (for totals) > Columns (few periods).
- Point-in-time comparison: Bar/Column. Horizontal bars for long labels or many items.
- Distribution: Histogram (shape), Box plot (compare groups), Strip/dot plot (small samples).
- Relationship: Scatter; add color/size for additional variables with caution.
- Part-to-whole: 100% Stacked Bar (≤5 parts), Treemap (many parts). Avoid pies for many slices.
- Ranking: Sorted bars; consider Pareto (bars + cumulative line).
- Composition over time: Stacked area for shares; consider small multiples if many series.
- Matrix patterns: Heatmap with clear legend and ordering.
Worked examples
Example 1 — Monthly signups by channel (trend + segments)
Question: How did signups change over 12 months, and which channels drove growth?
- Chart: Line chart for total trend. For channels, use small multiples of line charts or a stacked area if channels are few (≤4).
- Why: Lines show continuity over time; small multiples prevent stacking distortions.
- Tweaks: Same y-axis across small multiples; annotate campaign launch dates.
Example 2 — Revenue by product last quarter (point-in-time comparison)
- Chart: Horizontal bar chart, sorted descending.
- Why: Bars encode length best for categorical comparison; horizontal fits long product names.
- Tweaks: Show data labels for exact values; emphasize top 3 with subtle color.
Example 3 — Delivery time variability across 3 warehouses (distribution)
- Chart: Side-by-side box plots.
- Why: Box plots compare median, spread, and outliers concisely across groups.
- Tweaks: Add a dashed reference line for SLA (e.g., 48h).
Example 4 — Does ad spend drive revenue? (relationship)
- Chart: Scatter plot (x = spend, y = revenue), add a linear trendline.
- Why: Relationship + strength is visible; trendline summarizes correlation.
- Tweaks: Use one color; avoid over-encoding size unless meaningful (e.g., campaign duration).
Step-by-step: choose the chart
- Define the core question (trend, compare, distribution, relationship, part-to-whole?).
- Identify data types (time, category, numeric).
- Pick a chart family that fits the question.
- Decide granularity (daily vs. weekly, top N vs. all).
- Reduce clutter (limit colors, sort, show gridlines lightly).
- Add context (targets, baselines, annotations).
- Test clarity: can a teammate explain your chart in 10 seconds?
- [ ] I can state the question in one sentence.
- [ ] I chose a chart that encodes the right comparison.
- [ ] Axes and sorting support the message.
- [ ] Labels/legends are minimal but sufficient.
- [ ] Color is purposeful (not decorative).
Exercises (hands-on)
These mirror the exercises below. Do them here, then open the solution to compare.
Exercise 1 — Choose charts for 4 stakeholder questions
Given metrics for the last year: monthly active users, churn rate by plan, order value distribution, and marketing spend vs. signups.
- A. Show how MAU changed over time and highlight seasonality.
- B. Compare churn rate across 6 subscription plans (long plan names).
- C. Explain the spread of order values and identify outliers.
- D. Show if higher marketing spend relates to higher signups.
Deliverable: list the chart type(s) and 1–2 lines of justification each.
Common mistakes and self-check
- Mistake: Using pie charts with many slices. Fix: Use 100% stacked bar or treemap.
- Mistake: Bar charts that don’t start at zero. Fix: Bars must start at zero to avoid distortion.
- Mistake: Overlapping too many lines. Fix: Use small multiples or highlight one focal series.
- Mistake: Encoding too many variables (shape, color, size). Fix: Prioritize the main message; remove extras.
- Mistake: Unsuitable granularity (daily noise). Fix: Aggregate to weekly/monthly if the question is trend.
- Mistake: Unsorted categories. Fix: Sort to support the comparison (e.g., descending).
Self-check:
- [ ] Can someone answer the question without reading a paragraph of text?
- [ ] Is the primary comparison encoded by position/length (most accurate), not color hue?
- [ ] Would the message hold if data labels were hidden?
Practical projects
- Product performance dashboard: trend (last 13 months), top-10 products (bar), returns rate distribution (box).
- Marketing effectiveness: spend vs. conversions (scatter), channel ROI ranking (bar), weekly conversions (line) with annotations.
- Operations pulse: on-time delivery % (line), warehouse comparison (bars), delivery time distribution (box plots).
Who this is for
- BI Analysts and aspiring analysts who design dashboards and reports.
- Anyone who presents metrics to stakeholders and needs clear, accurate visuals.
Prerequisites
- Basic understanding of measures, dimensions, and time series.
- Comfort with reading common charts (bar, line, scatter, histogram).
Learning path
- Before: Data types and summarization (measures vs. dimensions).
- This lesson: Map questions to chart families; apply design tweaks.
- Next: Visual encoding best practices (color, scale, annotations) and dashboard layout.
Mini challenge
Design a 3-tile KPI view for the last quarter:
- Tile 1: Overall revenue trend with target line.
- Tile 2: Top 8 customers by revenue.
- Tile 3: Order value distribution.
Choose the chart for each tile, write one sentence explaining why, and note one design tweak that improves clarity.
Quick Test
Take the quick test to check understanding. The test is available to everyone; only logged-in users get saved progress.