Why this matters
As a Business Analyst, you turn raw numbers into decisions. Picking the right chart type makes your message obvious, reduces misinterpretation, and speeds up stakeholder alignment. You will use this skill when you: summarize KPIs for a weekly report, compare product performance across regions, explain a trend in churn, show distribution of delivery times, or highlight relationships between price and conversion.
- Clear visuals save time in meetings.
- Right charts reduce analysis errors.
- Stakeholders trust insights they can instantly read.
Concept explained simply
Every chart answers a question. First decide the question, then choose the chart.
- Comparison (which is bigger?) → bar/column, grouped bars.
- Trend over time (how is it changing?) → line (rate), area (part of whole over time), column (few time points).
- Distribution (what’s the spread?) → histogram, box plot, density.
- Composition (what are the parts?) → stacked bars, 100% stacked bars; use pies/donuts sparingly.
- Relationship (do two variables move together?) → scatter, bubble (adds size).
- Geography (where?) → choropleth (color-filled areas), symbol maps (bubbles on map).
Mental model: QCDR-G
Before you pick a chart, say QCDR-G out loud.
- Q = Question type (comparison, trend, distribution, composition, relationship, geography)
- C = Cardinality (how many categories/points?)
- D = Data type (categorical, numeric, time)
- R = Reading task (rank, compare, find pattern)
- G = Granularity (daily vs monthly, product vs category)
Then pick the chart that best matches the reading task with minimal cognitive load.
Quick decision helper
- Few categories (≤10) comparison → horizontal bar, sorted descending.
- Many categories → summary first (top N, others combined), then detail on demand.
- Monthly or daily metric → line (show trend); few periods (≤6) → columns okay.
- Part-to-whole with many slices → bar is better than pie. Pie only for 2–4 parts with clear differences.
- Skewed numeric data → histogram + median/quantiles; add box plot for compact summary.
- Suspected correlation → scatter with a light trend line.
Worked examples
Example 1: Monthly revenue trend
Question: Is revenue trending up?
Data: Monthly totals for 24 months.
Pick: Line chart. Add a 3-month moving average if noisy. Annotate major events (price change) with callouts.
Why not: Columns clutter across 24 months and hide rate of change.
Example 2: Support tickets by category
Question: Which categories drive the workload?
Data: 8 categories, counts last quarter.
Pick: Horizontal bar chart, sorted descending, data labels at the end.
Why not: Pie with 8 slices makes ranking hard.
Example 3: Trial-to-paid conversion vs discount
Question: Does discount relate to conversion?
Data: Discount %, conversion % by campaign.
Pick: Scatter plot. Add a faint trend line; label notable outliers.
Why not: Line charts imply time order that doesn’t exist.
Example 4: Delivery time variability
Question: What is the spread of delivery times?
Data: 5,000 orders, minutes to deliver.
Pick: Histogram to show distribution + vertical line for median. Optionally add a box plot for compact view by region.
Why not: Averages-only bar chart hides long delays.
Quick pattern guide (fast lookup)
- Bar (horizontal): compare categories, best for ranking.
- Column (vertical): few time periods or categories.
- Line: trends across many time points.
- Area: cumulative totals or part-to-whole over time (use sparingly, can obscure).
- Stacked bar: show composition + totals; 100% stacked for share, not magnitude.
- Pie/Donut: 2–4 parts with clear differences; avoid otherwise.
- Histogram: numeric distribution; choose sensible bin size.
- Box plot: distribution summary (median, quartiles, outliers).
- Scatter: relationship between two numeric variables.
- Heatmap: many categories by time; look for hotspots.
How to choose in practice
- Write the one-sentence question the chart must answer.
- Identify data types (time, category, numeric).
- Select chart aligned with the reading task (rank, trend, spread, part-to-whole, relationship).
- Reduce clutter: limit colors, sort, add direct labels, remove extra gridlines.
- Test with a colleague: ask what they see in 5 seconds.
Minimal styling checklist
- Titles state the takeaway, not just the metric.
- Axes start at zero for bars; for lines, zero may be optional if focus is change.
- Labels use consistent formats (%, currency).
- Color carries meaning (status, segment), not decoration.
- Annotations explain spikes/dips.
Common mistakes and how to self-check
- Using pies for many categories. Fix: switch to a sorted bar chart.
- Too many colors. Fix: use one highlight color; gray for context.
- Wrong axis start for bars. Fix: start at zero to avoid exaggeration.
- Overplotting in scatter. Fix: add transparency or bin points (hexbin), or sample with a note.
- Misleading time spacing. Fix: ensure equal time intervals on the x-axis.
Self-check prompt
Ask: Can a new viewer state the main insight in 5 seconds? If not, simplify the chart, sharpen the title, or choose a better type.
Who this is for
- Business Analysts preparing stakeholder dashboards and reports.
- New analysts learning to make data readable.
- Anyone responsible for decision-ready visuals.
Prerequisites
- Basic understanding of metrics and dimensions.
- Comfort with spreadsheets or BI tools (e.g., creating simple charts).
- Know your audience and the decision they need to make.
Learning path
- Identify your question type using QCDR-G.
- Pick a chart from the quick pattern guide.
- Draft a minimal, readable chart.
- Annotate with one key takeaway.
- Validate with a peer in 5 seconds.
Practical projects
- Weekly KPI board: churn, revenue, active users. Include one trend and one comparison chart.
- Support workload review: rank top categories; show composition by channel using 100% stacked bars.
- Delivery improvement deck: histogram of times; annotate median and 90th percentile; add regional box plots.
Exercises
Do these now. They mirror the exercises below and build real-world intuition.
Exercise 1: Map questions to charts
Goal: Choose chart types that fit the question and explain why.
- Scenario A: Quarterly revenue for last 8 quarters; stakeholders want the trend.
- Scenario B: NPS by 10 product features; find the lowest performers.
- Scenario C: Conversion vs load time across 150 landing pages.
Checklist:
- State the question in one sentence.
- Pick a chart type and justify in one line.
- Add one improvement (sort, label, annotation).
Exercise 2: Fix a misleading chart
Goal: Redesign a flawed chart decision.
- Current: Pie chart with 9 segments for market share.
- Task: Propose a clearer alternative and how you would style it.
Checklist:
- Choose a new chart type.
- Explain why it’s clearer.
- List 3 styling choices to reduce cognitive load.
Mini challenge
You receive a CSV with: date, region, channel, orders, revenue, refund_rate. Create a 1-page visual summary to answer: Which region underperformed this quarter and why? Your constraints: one trend chart, one comparison chart, and one annotation. Write down the three chart types you’ll use and the specific titles.
Next steps
- Apply QCDR-G to your current dashboard and replace at least one chart.
- Create a chart style checklist you reuse in every report.
- Take the quick test below. Note: Anyone can take the test for free. Only logged-in learners will see saved progress.