chart-interpretation

name	Chart Interpretation
description	Read any chart or data visualization, extract insights, detect patterns and anomalies, check for bias, and construct a narrative.

Purpose

Decode any chart (screenshot, embedded image, slide visual) into actionable insights. Charts can inform or mislead. This skill teaches how to read them critically, extract what matters, and spot what's hidden or distorted.

Steps

1. Identify the chart type: bar, line, pie, scatter, area, combo, heatmap, treemap, funnel
2. Read the axes: What's measured? What units? What time range? Are axes truncated?
3. Decode the encodings: What do colors, sizes, positions, and shapes represent?
4. Find the story: What's the main takeaway? What trend or comparison is shown?
5. Check for bias: Is the visualization honest or misleading?
6. Extract the narrative: Write what this chart tells the audience

Chart Type Identification

Type	Recognizable Feature	Best For
Bar	Rectangular bars, categorical axis	Comparing categories
Line	Connected points over continuous axis	Trends over time
Pie/Donut	Circle divided into segments	Part-of-whole (max 5 slices)
Scatter	Dots in 2D space	Correlation between variables
Area	Filled region under a line	Volume/composition over time
Heatmap	Color grid	Density or intensity patterns
Funnel	Narrowing stages	Conversion or dropout rates

Critical Reading Checklist

Check	What to Look For
Y-axis start	Does it start at 0? Truncated axes exaggerate differences
Scale linearity	Are intervals equal? Non-linear scales distort patterns
Time range	Is it cherry-picked to show a favorable trend?
Missing context	Is there a baseline, target, or benchmark for comparison?
Color manipulation	Does red = bad and green = good, or is color arbitrary?
Sample size	Is n= shown? Small samples produce unreliable patterns
Dual axes	Do two y-axes create a misleading correlation?
3D effects	Do they distort proportions? (Almost always yes)

Insight Extraction Template

For any chart, answer these questions:

1. WHAT: What does this chart show? (One sentence)
2. SO WHAT: Why does this matter? What's the implication?
3. PATTERN: What trend, comparison, or anomaly stands out?
4. OUTLIER: Is anything unexpected or inconsistent?
5. NOW WHAT: What action does this data suggest?

Common Misleading Patterns

Pattern	How It Misleads	How to Detect
Truncated y-axis	Small differences look dramatic	Check if y-axis starts at 0
Cherry-picked time range	Shows only the favorable period	Ask: what happened before and after?
Cumulative vs. incremental	Cumulative lines always go up	Check if the metric is cumulative
Correlation as causation	Two lines moving together implies link	Look for common causes or coincidence
Pie chart with too many slices	Impossible to compare accurately	Count the slices; more than 5 is suspect

Output Format

• Chart type identification and encoding summary
• Key insight in one sentence (WHAT + SO WHAT)
• Bias or quality assessment
• Recommended actions or follow-up questions

Guidelines

• Always check the axes before drawing conclusions
• If a chart looks dramatic, check if the scale is honest
• "Correlation" is not "causation": always look for alternative explanations
• A good chart answers one question clearly; a bad chart confuses many
• When presenting a chart to others, state the takeaway explicitly instead of saying "as you can see"