How Two Everyday Charts Got It Wrong (and What We Can Learn)
- Michael Lee, MBA
- 1 day ago
- 4 min read
Introduction
I didn’t expect to learn a lesson in data integrity while scrolling LinkedIn.
But there it was—a clean, well-labeled chart comparing Malaysia and Singapore’s GDP from 1967 to 2023. Something about it felt... off. Then I noticed the bars. Malaysia’s 2023 GDP looked shorter than its 1967 number, even though the economy had grown more than a hundredfold. A classic case of “looks right, feels wrong.”
Then it happened again. A few days later, I spotted a donut chart from Singapore’s Department of Statistics showing population breakdowns. It looked official. Polished. Credible. But it was visually misleading in a different way.
These weren’t bad charts. In fact, they were based on good data. But they conveyed the wrong message—visually. And that’s the kind of thing we train our learners to detect and redesign at FYT Consulting.
Let me show you what I mean.
Example 1: The Singapore Population Donut That Looks Official—but Misleads
On the surface, this chart from SingStat is neat and colorful. It splits Singapore’s 6.04 million total population into Citizens, Permanent Residents, and Non-Residents. Then it throws in another label: “Residents,” which equals Citizens + PRs.
That’s where things get messy.
Donut charts, like pie charts, imply all the segments are distinct parts of a whole. But here, one slice overlaps with two others. “Residents” isn’t separate—it’s a total that already includes two other slices.
That means readers who scan quickly might think:
Non-Residents outnumber Citizens (they don’t).
“Residents” is a fourth, separate category (it’s not).
It’s misleading, even if unintentionally.
On closer inspection, the donut chart also suffers from serious layout flaws. The gold ring meant to represent “Residents” is misaligned—it wraps around part of the Non-Resident segment instead of highlighting Citizens + PRs. Even the pointer line is misleading. These visual mistakes distract from what could have been a clear and informative message and instead introduce doubt and misinterpretation.
This chart, published by the Department of Statistics Singapore (SingStat), appears in their official annual summary: Singapore in Figures: Population and Households. 📎 Source: https://www.singstat.gov.sg/publications/reference/singapore-in-figures/population-and-households
What’s a better approach?
Instead of using a donut chart—which is best for distinct, non-overlapping categories—we recommend using a 100% stacked bar chart to show the composition of Singapore’s population clearly.
Here's how it would work:
The full bar represents 100% of the population (6.04M).
It's broken into Citizens, Permanent Residents, and Non-Residents.
Each segment’s proportion is visualized accurately, and overlapping definitions like "Residents" (Citizens + PRs) can be noted as a label instead of drawn as a slice.
Alternatively, a grouped bar chart could show all four values (Citizens, PRs, Non-Residents, and Residents) side-by-side while clarifying in a footnote or caption that Residents = Citizens + PRs.
A third option is a hybrid stacked/grouped bar chart: display two bars—one for Non-Residents and one for Residents—with the Residents bar broken down into Citizens and PRs. This avoids overlap, clarifies grouping, and visually communicates the composition more truthfully.
Three redesigns that avoid overlap and clarify the relationships: stacked, grouped, and hybrid bar charts.
This approach removes confusion and helps the viewer understand the data structure without misleading implications.
At FYT, we often use these kinds of before-and-after examples in class. Learners are asked to critique the original and redesign it themselves—so they don’t just consume visuals, they construct better ones.
Example 2: Malaysia vs Singapore GDP (1967 vs 2023)
This chart, which circulated on LinkedIn and originated from SEA Rising using World Bank data, compared the GDP of Malaysia and Singapore across a 56-year span.
In 1967, Malaysia’s GDP was US$3.19B while Singapore’s was US$1.24B. By 2023, Singapore had surged to US$501B, overtaking Malaysia’s US$399B.
All good—until you look at the bars.
Malaysia’s 2023 bar appears shorter than its 1967 bar, even though the economy expanded more than 100-fold. It visually implies Malaysia’s economy shrank, which is clearly not the case. This is a textbook case of poor scaling.
What would be more honest?
Make sure bar heights are proportional to actual values. If that means earlier values are visually tiny in comparison—so be it. That reflects reality. Alternatively, we often use indexed line charts or growth multiplier visuals in our sessions to show performance trends more meaningfully.
These aren’t just stylistic tweaks—they shape how people interpret insights.
Chart source: SEA Rising (Facebook page)Data source: World Bank (Current US$ GDP)Accessed via: https://data.worldbank.org/indicator/NY.GDP.MKTP.CD?locations=MY-SGRetrieved: 2025
Key Data Visualization Principles (From These Examples)
🔹 Avoid overlapping categories in part-to-whole visuals
Donut and pie charts should only be used when categories are distinct. If values overlap (e.g., Residents = Citizens + PRs), use bars or stacked visuals instead.
🔹 Always scale visuals to match data
When one value is 100 times larger, it should look that way. Don't shrink or stretch bars for visual symmetry—truth first.
🔹 Label directly when possible
Floating labels with thin pointers create confusion. If your reader needs a legend to decode a simple visual, it's not simple enough.
🔹 Choose chart types based on your message
Are you showing trends, comparisons, or breakdowns? Let your intent guide your design—not the shape of the chart.
🔹 Clarity beats cleverness
A simple bar done right can out-communicate a sleek donut done wrong.
And ultimately, we believe the best way to learn data storytelling is through application—and sometimes,
the most teachable moments come from visuals we see every day.
Final Thought
Next time you see a chart—whether on LinkedIn, in a news report, or from a government website—take a second look. Ask yourself: Is this the right way to present this message? And if you’re the one presenting the data, remember: the best charts don’t just show numbers. They clarify meaning.
Comments