Lesson 10 of 16 · Level 3: Advanced
Data Visualization: How to Choose the Right Chart
- ⏱ 11 min read
- ❓ 4 review questions
- ✍️ Exercise with solution
- 📥 Downloadable template
In this lesson
- How to pick the right chart by first naming the question: comparison, trend, composition, distribution or relationship.
- The design rules that make bar and line charts honest and easy to read.
- How to use color with purpose: one highlight, gray for the rest, and accessible palettes.
- How to write chart titles that state the takeaway instead of describing the axes.
- The most common chart mistakes, including truncated axes and overloaded pie charts.
Recreate the charts in this lesson from the Summary sheet: it already has revenue by month, region, category and channel. Select a block and use Insert → Chart.
- sales-practice-data.xlsx · sheet Summary
A chart is an argument, not decoration. The right chart makes the main finding obvious in five seconds; the wrong one hides it, or worse, distorts it. The good news is that choosing well is a process, not a talent: decide what question the chart answers, pick the chart type that fits that question, then remove everything that doesn't help. This lesson walks through that process with charts built from the course dataset.
Start with the question, not the chart type
Most chart questions fall into a handful of families. Name the family first and the chart choice becomes almost automatic:
| The question is about… | Best charts | Example from our data |
|---|---|---|
| Comparing categories | Bar chart (horizontal for long labels) | Revenue by product |
| Change over time | Line chart (column chart for few periods) | Revenue by month |
| Parts of a whole | Stacked or 100% bar; donut with 2–4 slices | Revenue share by channel |
| Distribution | Histogram, box plot | Order values (Lesson 6) |
| Relationship between two numbers | Scatter plot | Units vs. revenue (Lesson 13) |
| Geography | Filled map (with care), or a sorted bar | Revenue by state |
| One key number | Big number (KPI card) with a comparison | 2025 revenue: $89,735 |
Bar charts
The workhorse. People compare lengths very accurately, so bars beat almost everything for comparing categories.
Rule: start the axis at zero and sort by value.Line charts
Show how a measure moves over continuous time: days, weeks, months.
Rule: time on the horizontal axis, oldest on the left; limit to a few lines.Pie and donut charts
Only for parts of a whole that add to 100%, with very few slices.
Rule: more than 4 slices or slices of similar size? Use a sorted bar instead.Scatter plots
Show whether two numeric variables move together, and reveal clusters and outliers.
Rule: color points by a category to expose hidden groups.1. Your manager wants to see how revenue moved across the 12 months of 2025. Which chart fits best?
Change over time calls for a line chart, read left to right in time order. A 12-slice pie can't show order or trend, and order IDs aren't a meaningful axis.
Change over time: the monthly revenue line
Notice the choices: the axis starts at $0, there are only a few light gridlines, both lines are labeled directly (no legend to decode) and the two peaks are annotated with their values. Showing the consumer line in gray gives context without competing with the main story.
Comparing categories: the sorted bar chart
Here's revenue by product, sorted from largest to smallest. Sorting turns a list into a ranking the reader understands instantly:
Three design decisions make this chart work: horizontal bars (product names are long and stay readable), sorted values and one highlight color on the bar the title talks about, with everything else in gray. The same data as a 10-slice pie chart would be almost unreadable: nobody can tell a 6.2% slice from a 5.4% slice.
Select the product and revenue columns of a summary table (or a pivot), then Insert → Recommended Charts → Clustered Bar in Excel, or Insert → Chart in Sheets and choose "Bar chart" in the Chart editor. Sort the source table by revenue (largest first) before charting. In Excel, a bar chart lists categories bottom to top, so if the largest bar ends up at the bottom, check "Categories in reverse order" in the vertical axis options.
Honest axes: why bars must start at zero
The length of a bar is its value. If the axis starts above zero, lengths lie. Take Northeast revenue ($20,786) and South revenue ($18,684): the Northeast is about 11% higher. On a bar chart whose axis starts at $18,000, the Northeast bar would be 2,786 units long and the South's 684, so it would look about 4 times taller.
Line charts encode values by position, not length, so a line chart's axis may start above zero to show variation, for example a stock price between $180 and $195. Just label the axis clearly. Bars and areas, never.
2. A bar chart compares two regions with revenue of $20,786 and $18,684, and its axis starts at $18,000. How much taller does the first bar appear?
The visible bar lengths are 20,786 − 18,000 = 2,786 and 18,684 − 18,000 = 684, and 2,786 ÷ 684 ≈ 4.1. The real ratio is 20,786 ÷ 18,684 ≈ 1.11. Truncated bar axes exaggerate differences.
Use color with intention
- Gray is your default. Use one strong color to highlight what the title says. If everything is colorful, nothing stands out.
- Same thing, same color. If Wholesale is orange in one chart, keep it orange on every chart and slide.
- Sequential vs. categorical. Use shades of one color for ordered values (low to high), distinct hues for unrelated categories.
- Accessibility. Red-green color blindness is common, especially among men. Don't rely on red vs. green alone; add labels, use blue vs. orange or vary lightness.
Titles that say something
A descriptive title tells the reader what the chart is. A takeaway title tells them what it means. In a report or presentation, use takeaway titles:
| Descriptive title | Takeaway title |
|---|---|
| Revenue by product | Smart watches brought in 26.2% of 2025 revenue |
| Monthly revenue 2025 | Two wholesale-driven spikes dominate 2025; consumer sales were steady |
| Revenue by region | The West generated 40.8% of revenue, more than any two other regions combined |
That last title is checkable: Northeast plus South is $39,470, which is more than the West's $36,626, so the claim would be false. The correct version: "The West generated 40.8% of revenue, about 1.8 times the next region" ($36,626 ÷ $20,786 ≈ 1.76). Takeaway titles are powerful, which is exactly why you must verify them.
- Quantity
- Dozens, quick and rough.
- Goal
- Find patterns, errors and questions.
- Quantity
- One or two per key message.
- Goal
- Make one finding obvious and memorable.
3. You need to show the revenue share of 10 products. What's the best option?
Ten slices are too many for a pie, and 3D distorts slice sizes further. A sorted bar chart shows the ranking and the exact values. Lines imply a continuous sequence, which products don't have.
Chart mistakes to avoid
3D effects, shadows and gradients.
Flat, simple shapes. 3D distorts sizes and adds nothing.
A legend the reader must match to colors.
Label lines and bars directly.
Two different measures on two Y axes.
Use two charts side by side; dual axes invite false conclusions.
Bars in alphabetical or random order.
Sort by value, unless the categories have a natural order (months, sizes).
Ten colors because the software offered them.
Gray plus one highlight color.
4. Which title is best for a slide that shows wholesale's share of orders and of revenue?
The third title states the finding with numbers, so the audience knows what to look for before reading the chart. The others describe the chart without saying why it matters.
Practice exercise
For each question, choose a chart type and write a takeaway title using these figures from the practice data:
- How did online + retail revenue change over the year? (First half: $11,679. Second half: $16,577.)
- Which states bring in the most revenue? (California $15,467, Arizona $9,776, Texas $8,489, New York $8,440, Washington $8,262.)
- How does wholesale's share of orders compare with its share of revenue? (37 of 200 orders; $61,479 of $89,735.)
See solution
- Line chart of monthly online + retail revenue. Title: "Consumer sales grew 42% from the first half to the second half of 2025" ($16,577 ÷ $11,679 ≈ 1.42).
- Sorted horizontal bar chart of the top 5 states, California highlighted. Title: "California alone brought in $15,467, about 1.6 times second-place Arizona" ($15,467 ÷ $9,776 ≈ 1.58). A map would make small, high-revenue states hard to compare.
- Two 100% stacked bars (orders and revenue), wholesale highlighted. Wholesale is 37 ÷ 200 = 18.5% of orders and $61,479 ÷ $89,735 = 68.5% of revenue. Title: "Wholesale is fewer than 1 in 5 orders but more than two-thirds of revenue."
Summary
- Name the question family first: comparison, trend, composition, distribution, relationship.
- Bars for comparing categories (sorted, starting at zero); lines for change over time.
- Use pies only for 2–4 parts of a whole; prefer sorted bars otherwise.
- Gray by default, one highlight color, direct labels, no 3D.
- Write takeaway titles, and verify every claim in them against the data.
Next lesson: Dashboards with Power BI and Looker Studio. You'll combine KPIs, a trend and filters into an interactive dashboard in two free tools.
Create your account in 30 seconds, mark your lessons and download your certificate at the end.