You have just finished training a logistic-regression model that flags potentially fraudulent B2B invoices. For next week's 10-minute board meeting, the CFO wants one slide that instantly shows how many legitimate invoices would be held for manual review (false positives) and how many fraudulent invoices the model might miss (false negatives). Several board members are color-blind and have little time for technical explanations. Which visualization and design choice will best satisfy these communication requirements?
A 3-D stacked pie chart that uses red and green slices to depict true positives, false positives, true negatives, and false negatives.
An annotated confusion-matrix heatmap that uses a color-blind-safe blue-orange palette and displays the four cell counts in large text.
An ROC curve showing the area under the curve (AUC) with an interactive threshold slider.
A scatter plot of predicted fraud probability versus invoice amount, colored by the model's predicted class labels.
A confusion-matrix heatmap uses the same 2×2 layout as the underlying error counts, so non-technical executives can immediately map each cell to a real-world outcome (true/false and positive/negative). Showing the counts as large annotations removes the need for the audience to estimate values from color alone, and selecting a blue-orange (or other color-blind-safe) palette ensures that board members with red-green deficiencies see clear contrast. ROC curves and probability scatterplots require an understanding of thresholds or continuous scores that most executives do not possess. A 3-D pie chart exaggerates areas, makes comparison difficult, and relies on red/green hues that many viewers cannot distinguish. Therefore, the annotated, color-blind-friendly confusion-matrix heatmap is the most effective and accessible choice.
Ask Bash
Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.
What is a confusion matrix and why is it useful?
Open an interactive chat with Bash
Why is a color-blind-safe palette important in data visualization?
Open an interactive chat with Bash
Why are ROC curves or 3-D pie charts unsuitable for executive audiences?