These 12 CBDA data visualization questions test whether you can pick the right chart for a business goal, read technical plots described in words or small tables, and spot the anti-patterns the Guide warns about. Four questions match a chart to composition, flow or distribution data, four read boxplots, ROC curves, residual and autocorrelation plots, three fix flawed charts, and one translates a clustering plot for an executive. Answers and Guide references follow the questions.
How this cbda data visualization questions set is built
Most questions sit in Interpret and Report Results (Domain 4), task 2.4.5; the technical plots also touch Analyze Data (Domain 3). For a wider Domain 4 drill, try our Interpret and Report Results quiz.
All 12 are original scenarios, not recalled exam items, each grounded in a named section of the Guide to Business Data Analytics. Difficulty sits slightly above typical question banks: you get a business goal or a described plot and choose the best action.
The Guide ideas behind the set:
- 3.2 Business Visualizations. Visuals for stakeholders, chosen by what the data must show: comparison (bar, line, butterfly), composition (stacked bar, waterfall, area), distribution (line, histogram, scatter) and relationships (scatter, heat map, chord, Venn). Purpose is set by two questions, declarative or exploratory and conceptual or data-driven. Guidelines: know the problem and the stakeholders, stay unbiased, and prefer function over form.
- 3.18 Technical Visualizations. Autocorrelation, Box-Cox, contour, correlation map, PCA, residual, ROC and silhouette plots help data scientists confirm modelling assumptions; analysts translate them for stakeholders.
- 3.11 Exploratory Data Analysis. Univariate plots such as histograms show a variable's shape; among bivariate plots, boxplots are singled out for spotting outliers.
- 2.4.5 Document and Communicate Findings. One insight per graph, 2D over 3D, no secondary axis, horizontal bars instead of hard-to-read pies.
| Sub-topic | Questions |
|---|---|
| Choosing a chart for composition, flow or distribution | 1, 2, 3, 4 |
| Reading boxplots, ROC, residual and autocorrelation plots | 5, 6, 7, 12 |
| Fixing visualization anti-patterns | 8, 9, 10 |
| Translating a technical visual for a stakeholder | 11 |
Questions 2 and 5 need arithmetic; question 10 is the only "Choose two" item. For the statistics behind these plots, see our CBDA statistics questions.
CBDA exam at a glance
| Fact | Detail |
|---|---|
| Format | 75 multiple-choice, scenario-based questions |
| Time | 2 hours |
| Domain weights | Identify the Research Questions 20%, Source Data 15%, Analyze Data 16%, Interpret and Report Results 20%, Use Results to Influence Business Decision Making 20%, Guide Organization-level Strategy for Business Analytics 9% |
| Exam fee | US$405, or US$250 for IIBA members; first year of membership included with exam purchase |
| Passing score | Not published by IIBA |
Last verified: October 2026
Practice Questions
Question 1
Northbridge Health earns revenue from hospitals, outpatient clinics and telehealth. The CFO wants one slide showing total revenue for 12 quarters and each service line's part of it, so the board sees telehealth grow.
Which visual BEST fits this request?
- A. A scatter plot of telehealth revenue against hospital revenue
- B. A stacked area chart with one band per service line
- C. A histogram of quarterly revenue values for all lines
- D. A Venn diagram of overlap between the service lines
Question 2
Lumen Freight's operating profit was 4.2 million last year. Drivers of change: price increase +1.1 million, fuel cost -0.8 million, wages -0.3 million, route optimisation +0.4 million. The CFO wants one visual showing how last year's figure became this year's, driver by driver.
Which visual and final value should the analyst present?
- A. A waterfall chart from 4.2 million ending at 5.2 million
- B. A stacked bar splitting this year's 4.6 million by driver
- C. A pie chart of each driver's share of the total change
- D. A waterfall chart from 4.2 million ending at 4.6 million
Question 3
Harbor Grocer has delivery times for 38,000 online orders. The operations manager suspects same-day and scheduled orders form two separate clusters of delivery times. The analyst has not yet looked at the data.
Which visual should the analyst use FIRST to examine this?
- A. A histogram of delivery times with consistent bins
- B. A pie chart of orders by delivery-time band
- C. A Venn diagram of same-day and scheduled orders
- D. A chord diagram of orders between depots and zones
Question 4
Aster Mobile has six price plans, and thousands of customers switch plans each month. The commercial director wants one view of how many customers moved between every pair of plans last year, not just each plan's net change.
Which visual BEST meets the director's need?
- A. A waterfall chart of net subscriber change per plan
- B. A run-sequence plot of total migrations per month
- C. A chord diagram with ribbons sized by customers moved
- D. A histogram of plan changes made per customer
Question 5
Oakline Logistics summarises boxplots of order pick times (minutes) at three sites:
| Site | Q1 | Median | Q3 | Points beyond whiskers |
|---|---|---|---|---|
| North | 6 | 8 | 10 | 25, 27 |
| East | 7 | 9 | 16 | none |
| West | 5 | 7 | 8 | none |
Which conclusion is BEST supported by these boxplots?
- A. North has the most variable typical times, given its two outliers
- B. East has the widest middle half, an interquartile range of 9 minutes
- C. West is the slowest site, since its box sits lowest on the axis
- D. North's average pick time is 8 minutes, read from its box
Question 6
Crestline Bank built two card-fraud classifiers. On the same validation data, Model A's area under the ROC curve is 0.71 and Model B's is 0.86. The fraud team can review only a limited number of alerts a day, so false alarms are costly.
What should the analyst conclude?
- A. Prefer A, since a smaller area means fewer false alarms
- B. Prefer B, since 0.86 means 86% of live predictions are correct
- C. Prefer B and set the threshold with the highest true positive rate
- D. Prefer B, then pick a threshold that keeps false alarms workable
Question 7
Orchard Homes predicts monthly rent with linear regression. On the residual plot, errors for cheaper units sit in a tight band around zero, but the spread widens steadily as predicted rent rises, forming a fan.
What does this pattern MOST likely indicate?
- A. Heteroskedasticity, possibly from subgroups that behave differently
- B. A non-linear relationship the linear model has missed
- C. Seasonality in rents that an autocorrelation plot would confirm
- D. A well-fitted model, because residuals are centred on zero
Question 8
Pinewood Outdoor's board deck has one chart with bars for quarterly revenue and a profit-margin line on a secondary y-axis. Two directors later said margin had collapsed; they had read the line against the revenue axis.
What is the BEST revision for the next deck?
- A. Render bars and line in 3D so the series separate visually
- B. Add a third axis for unit volume to give full context
- C. Split it into two charts, one for revenue and one for margin
- D. Replace it with a pie chart of annual revenue and profit
Halfway there. If these six felt manageable, create a free account to keep practising. You get more CBDA questions, including the full 75-question CBDA exam simulator with exam and practice modes. certificate.tips is our product.
Question 9
After a checkout redesign, Vela Travel's payment-step conversion rose from 95.1% to 96.0%. The product lead asks the analyst to start the bar chart's y-axis at 94.8% so the "after" bar looks four times taller.
What should the analyst do?
- A. Use the 94.8% baseline, since the values shown are still true
- B. Use a scale that does not exaggerate and state the 0.9-point rise
- C. Switch to 3D bars so the gap shows without a new baseline
- D. Remove the y-axis labels so readers focus on the visual gap
Question 10
Fernway Retail's sales dashboard shows revenue share across 11 product categories as a 3D pie chart. Most slices are 6% to 11%, and users struggle to rank categories.
Which TWO changes BEST follow visualization best practice? (Choose two.)
- A. Replace the pie with a horizontal bar chart sorted by share
- B. Add a secondary y-axis showing unit volume on the same chart
- C. Keep the pie but explode each slice so categories stand apart
- D. Split the largest categories into more slices for detail
- E. Use a 2D form with clear labels and colours, not 3D depth
Question 11
Brightleaf Insurance's data scientist split customers into eight clusters. On the silhouette plot, five clusters are wide and long; clusters 2 and 6 are narrow and start in the negative range. The marketing VP, who has no statistics background, asks what this means for campaigns.
What is the analyst's BEST response?
- A. Show the VP the plot and explain how silhouette scores work
- B. Recommend eight campaigns, one per cluster the model produced
- C. Ask the data scientist to replace the plot with a ROC curve
- D. Explain that two segments are not distinct and fewer may work
Question 12
Saltmarsh Utilities has four years of weekly water demand. Before choosing a forecasting model, the data scientist wants to check whether demand depends on earlier weeks and whether a yearly cycle exists.
Which visual should the analyst expect the data scientist to use?
- A. A correlation map of pairwise weather-variable correlations
- B. A silhouette plot of how well weeks group into clusters
- C. An autocorrelation plot of demand correlation at each lag
- D. A Box-Cox plot to find a transformation that normalises demand
Ready for the full-length version? The 12 questions above are a sample. Sign up and create a free account to unlock the full 75-question CBDA exam simulator, timed exam mode for a realistic run and practice mode with explanations after each answer.
Answer Key and Explanations
1. Correct answer: B. A whole split into parts and tracked over time is composition data. The Guide names stacked bars, waterfalls and area charts for composition; a stacked area chart keeps the time axis, so the telehealth band visibly widens. Source: Guide to Business Data Analytics, 3.2 Business Visualizations Why the others are wrong: A shows a relationship, not parts of a total; C loses the time order; D shows set overlap, which revenue lines do not have.
2. Correct answer: D. 4.2 + 1.1 - 0.8 - 0.3 + 0.4 = 4.6 million. The waterfall is one of the Guide's composition visuals and shows a start value turning into an end value through additions and subtractions. Source: Guide to Business Data Analytics, 3.2 Business Visualizations Why the others are wrong: A adds the wage increase instead of subtracting it; B cannot show negative drivers; C cannot show negative contributions.
3. Correct answer: A. Two suspected clusters is a question about one variable's distribution shape. The Guide lists histograms among univariate EDA plots that show frequency and shape, and a bimodal pattern appears as two peaks. Source: Guide to Business Data Analytics, 3.11 Exploratory Data Analysis Why the others are wrong: B hides shape inside arbitrary bands; C shows set membership; D shows flows between locations.
4. Correct answer: C. Movements between every pair of plans are relationships among many categories. The Guide names chords among visuals for relationships, and ribbon width can carry the customer count. Source: Guide to Business Data Analytics, 3.2 Business Visualizations Why the others are wrong: A shows only the net change the director rejected; B loses the pairs; D shows a per-customer distribution.
5. Correct answer: B. The box spans Q1 to Q3, so its length is the interquartile range: East 16 - 7 = 9 minutes, North 4, West 3. The Guide notes boxplots are good for identifying outliers, which is a separate reading from the spread of the box. Source: Guide to Business Data Analytics, 3.11 Exploratory Data Analysis Why the others are wrong: A confuses extreme points with typical spread; C reverses the axis, as a lower box means faster picks; D treats the median as a mean.
6. Correct answer: D. The Guide says more area under the ROC curve means a better classifier, and that moving the threshold trades true positive rate against false positive rate until the balance suits the problem. Fraud detection is its own example. Source: Guide to Business Data Analytics, 3.18 Technical Visualizations Why the others are wrong: A reverses the meaning of the area; B treats AUC as live accuracy; C ignores the cost of false alarms.
7. Correct answer: A. A residual spread that changes across predictions means non-constant variance. The Guide calls this heteroskedasticity and notes it may signal subgroups with different attributes, which can change the model used. Source: Guide to Business Data Analytics, 3.18 Technical Visualizations Why the others are wrong: B matches a curved pattern such as the Guide's inverted U, not a fan; C applies a time-series plot to a cross-sectional problem; D ignores that the pattern shows fit.
8. Correct answer: C. The Guide warns that a bar graph and a line superimposed on a secondary y-axis cause confusion, using quarterly revenue and a profit trend as its example. It also recommends one graph per insight. Source: Guide to Business Data Analytics, 2.4.5 Document and Communicate Findings from Completed Analysis Why the others are wrong: A adds depth, which is hard to read; B adds another axis; D drops the quarterly trend.
9. Correct answer: B. From a 94.8% baseline the bars would be 0.3 and 1.2 points tall, a fourfold visual gap for a 0.9-point gain. The Guide's guideline to "present unbiased analysis" shows one data set on two y-axis scales and warns the stretched one can bias the audience. Source: Guide to Business Data Analytics, 3.2 Business Visualizations Why the others are wrong: A ignores that true labels do not prevent a false impression; C swaps one distortion for another; D removes context.
10. Correct answers: A and E. The Guide says a pie chart that forces readers to compare arc lengths and angles can be replaced with a horizontal bar chart, and that 2D graphs with suitable colours and labels often beat 3D because depth is hard to judge. Source: Guide to Business Data Analytics, 2.4.5 Document and Communicate Findings from Completed Analysis Why the others are wrong: B adds a secondary axis, which the same list advises against; C keeps the angle problem; D adds slices to a crowded chart.
11. Correct answer: D. In the Guide's silhouette example, narrow clusters and clusters starting in the negative range can be removed, leaving fewer segments. The Guide also expects analysts to translate technical visuals for stakeholders rather than pass them on raw. Source: Guide to Business Data Analytics, 3.18 Technical Visualizations Why the others are wrong: A hands a business leader an untranslated artefact; B ignores cluster quality; C swaps in a plot built for classifiers.
12. Correct answer: C. Autocorrelation plots test whether outcomes correlate with their own past values and matter most in time series. The Guide notes values near zero suggest randomness, while periodicity points to seasonality. Source: Guide to Business Data Analytics, 3.18 Technical Visualizations Why the others are wrong: A checks dependence between different variables; B evaluates clustering; D addresses skewed distributions.
How to score this set
Count one point per question; question 10 scores only if both choices are right.
- 11–12 correct. Solid chart judgement. Move on to timed practice.
- 8–10 correct. Review misses by sub-topic: questions 5–7 and 12 point to the technical plots in 3.18; questions 8–10 point to the 2.4.5 best-practice list.
- Below 8. Re-read the four data intents in 3.2 and the EDA visuals in 3.11, then work through our Analyze Data questions to link the plots to the analysis behind them.
To test yourself under exam conditions, create a free account and open the CBDA exam simulator, which has 75 questions with exam and practice modes.
What candidates say about data visualization on the CBDA
Candidates describe an exam that rewards reasoning over recall. The certificate.tips author who passed, writing in our CBDA certification overview, flagged about 20 of 75 questions and found the exam harder than question banks, with more statistical reasoning and analytical decisions than definitions. Mufaddal Haidermota, who earned the CBDA in December 2022, recommends per-domain mock questions, full-length tests, mind maps and cheat sheets. A Techcanvass candidate story describes a developer who relied on mind maps and suggests two to three hours of daily study. Trainer LN Mishra of Adaptive US advises daily practice, timed full-length mocks and a close review of every wrong answer, which suits chart questions where the distractors are plausible.
FAQ
Which chart types does the CBDA exam expect you to know?
Section 3.2 of the Guide names business charts by data intent, from bar and line charts to waterfalls, histograms, heat maps, chord and Venn diagrams. Section 3.18 adds technical plots such as ROC, residual, autocorrelation and silhouette plots, and 3.11 adds boxplots. Questions ask which visual fits a goal, not how to draw it.
How do I choose between a business and a technical visualization on the exam?
Ask who the visual is for. The Guide describes business visualizations as tools to communicate insights to stakeholders, while technical visualizations help data scientists understand data and confirm modelling assumptions. If the scenario involves an executive or a board, a business visual or a plain-language translation is usually the right answer.
What visualization mistakes does the Guide warn against?
Section 2.4.5 warns against 3D charts, bar and line graphs on a secondary axis, and pie charts that force readers to compare angles. Section 3.2 adds biased y-axis scales and complexity for its own sake.
Do I need to calculate statistics to answer chart questions?
Some questions need light arithmetic, such as an interquartile range from a boxplot table or the end value of a waterfall. You should know what a box, an area under a curve or a residual pattern represents.
Sources
- Certification in Business Data Analytics (CBDA) — IIBA, checked October 2026
- Certification Fees — IIBA, checked October 2026
- CBDA Certification Handbook — IIBA, checked October 2026
- Box Plot (e-Handbook of Statistical Methods, 1.3.3.7) — NIST/SEMATECH, accessed October 2026
- Autocorrelation Plot (e-Handbook of Statistical Methods, 1.3.3.1) — NIST/SEMATECH, accessed October 2026
- Business Data Analytics Certification — Mufaddal Haidermota, Medium, January 2023
- CBDA Certification Journey: Ivan — Techcanvass, 2024
- Mastering the CBDA Exam — LN Mishra, Adaptive US, October 2024