Graph Logic
Graph Logic is the principle of choosing the most appropriate type of graph based on the data
type, relationship, and communication goal so that information is presented clearly, accurately,
and without misleading the viewer.
2. Objectives of Graph Logic
• To represent data clearly and effectively
• To avoid misinterpretation and distortion
• To match data types with correct visual forms
• To support quick decision-making
3. Data Types and Suitable Graphs
A. Categorical (Qualitative) Data
Used when data represents categories or groups.
Suitable graphs:
• Bar Chart
• Pie Chart
Example:
Gender distribution, product categories
B. Numerical (Quantitative) Data
1. Discrete Data
Countable numeric values.
Suitable graphs:
• Bar Chart
Example:
Number of students per class
2. Continuous Data
Values in a range.
Suitable graphs:
• Line Chart
• Histogram
Example:
Temperature, height, time
C. Time-Series Data
Data collected over time.
Suitable graph:
• Line Chart
Example:
Daily air quality index
D. Relationship Between Variables
To observe correlation.
Suitable graph:
• Scatter Plot
Example:
Study hours vs exam score
4. Common Graph Types and Their Logic
1. Bar Chart
• Compares different categories
• Bars should have equal width
• Y-axis should start from zero
Used for: Comparisons
2. Line Chart
• Shows trends and changes over time
• Points connected by lines
Used for: Trends
3. Pie Chart
• Represents parts of a whole
• Total must equal 100%
• Limited categories (max 5–6)
Used for: Proportions
4. Histogram
• Shows frequency distribution
• Bars touch each other
Used for: Continuous data distribution
5. Scatter Plot
• Displays relationship between two variables
• Identifies correlation (positive/negative)
Used for: Relationships
Data Type Description Example Common Charts
Nominal Gender, Product type, Bar chart, Pie chart,
Categories with no order
(Categorical) Country Column chart
Ratings (Poor, Good,
Categories with Bar chart, Column chart,
Ordinal Excellent), Education
meaningful order Heatmap
level
Numeric data with
Line chart, Scatter plot,
Interval meaningful differences, no Temperature (°C), Year
Area chart
true zero
Numeric data with Age, Height, Sales Histogram, Line chart,
Ratio
meaningful zero revenue Scatter plot
Only two possible
Binary Yes/No, True/False Pie chart, Bar chart
categories
Stock prices over
Line chart, Area chart,
Time series Data collected over time months, Daily
Sparkline
temperature
Spatial / Population by district, Maps, Choropleth maps,
Data tied to locations
Geographical Earthquake locations Bubble maps
Words or sentences for Customer reviews, Word clouds, Bar charts
Text / String
analysis Tweets (frequency), Heatmaps
Hierarchical / Data with nested Company structure, File Tree maps, Sunburst
Tree relationships directories charts, Dendrograms
Network / Connections between Social media friends, Network graphs, Force-
Relational entities Flight routes directed graphs
1. Clarity
Definition:
Clarity means the visualization communicates information in the simplest, most understandable
way. The viewer should immediately grasp the main message without confusion.
Why it matters:
• Confusing visuals can mislead or frustrate viewers.
• Clear visuals reduce cognitive load and help faster decision-making.
How to achieve clarity:
• Use simple chart types appropriate to the data.
• Avoid 3D effects, unnecessary decorations, or excessive text.
• Label axes, data points, and categories clearly.
Example:
• Clear: A bar chart showing monthly sales. Each bar labeled, Y-axis starts at zero.
• Unclear: A 3D pie chart with overlapping slices and no labels.
Common mistakes:
• Overcrowded charts.
• Using chart types that don’t fit the data.
2. Accuracy
Definition:
Accuracy ensures that the visualization represents the data truthfully without exaggeration,
distortion, or misrepresentation.
Why it matters:
• Misleading charts can cause wrong conclusions and decisions.
• Accuracy builds trust in reports and dashboards.
How to achieve accuracy:
• Always use correct scales.
• Ensure that proportions, angles, and lengths reflect true values.
• Avoid truncated axes in bar charts unless clearly noted.
Example:
• Accurate: Pie chart where slices match actual percentages.
• Misleading: Bar chart starting Y-axis at 50 to exaggerate differences.
Common mistakes:
• Manipulating axes to overstate trends.
• Using visual elements that distort size or proportion.
3. Efficiency
Definition:
Efficiency is about conveying information quickly and effectively. A good visualization
communicates insights with minimal effort from the viewer.
Why it matters:
• Busy professionals need to make decisions fast.
• Visuals should minimize interpretation time.
How to achieve efficiency:
• Highlight key trends or patterns.
• Reduce unnecessary details.
• Use simple colors and fonts.
Example:
• A heatmap showing customer activity across regions allows instant identification of high
and low activity zones.
Common mistakes:
• Including too much data in one chart.
• Using fancy effects that slow down understanding.
4. Simplicity
Definition:
Simplicity is removing non-essential elements from a visualization. Less is more.
Why it matters:
• Simple visuals are easier to interpret.
• Reduces confusion and enhances focus on the key message.
How to achieve simplicity:
• Use minimal gridlines, labels, and legends.
• Avoid 3D charts unless necessary.
• Keep text and annotations concise.
Example:
• Simple line chart for stock prices over time instead of a 3D chart with gradient colors.
Common mistakes:
• Over-decorating charts with patterns, textures, or colors.
5. Consistency
Definition:
Consistency means using the same design conventions across multiple visualizations to aid
comparison and understanding.
Why it matters:
• Helps users quickly interpret charts without re-learning symbols or colors.
• Makes dashboards and reports professional and coherent.
How to achieve consistency:
• Use consistent colors for the same categories.
• Keep chart types uniform for similar comparisons.
• Maintain uniform scales for similar metrics.
Example:
• Red always represents losses, green always represents gains in all charts.
Common mistakes:
• Changing colors for the same category in different charts.
• Using multiple scales unnecessarily.
6. Visual Hierarchy
Definition:
Visual hierarchy guides the viewer’s eye to the most important information first using size, color,
position, or emphasis.
Why it matters:
• Highlights key insights.
• Prevents viewers from missing important information.
How to achieve visual hierarchy:
• Use bold or bright colors for important elements.
• Place key data points in prominent positions.
• Use size or thickness to emphasize main trends.
Example:
• In a sales report, highlight the top-selling product bar in bright blue while others are gray.
Common mistakes:
• All elements look equally important, making it hard to identify the main message.
7. Appropriate Scale
Definition:
The scale of axes or color gradients must reflect the data correctly to avoid misleading
impressions.
Why it matters:
• Wrong scales can exaggerate or downplay trends.
• Correct scale ensures accurate comparisons.
How to achieve appropriate scale:
• Start Y-axis at zero for bar charts.
• Use proportional sizes for bubble charts.
• Choose color scales that reflect true differences.
Example:
• Bar chart showing sales starts Y-axis at zero to accurately reflect differences.
Common mistakes:
• Truncated axes to exaggerate differences.
• Uneven or inconsistent intervals on axes.
8. Effective Use of Color
Definition:
Color should be used to convey meaning, group categories, or show intensity.
Why it matters:
• Color helps differentiate categories or highlight trends.
• Poor color use can confuse or mislead viewers.
How to achieve effective color usage:
• Use sequential colors for intensity (light to dark).
• Use distinct colors for categories.
• Avoid too many colors in one chart.
• Consider color-blind-friendly palettes.
Example:
• Heatmap using a gradient from light yellow (low) to dark red (high) to show customer
density.
Common mistakes:
• Random colors with no meaning.
• Using colors that are hard to distinguish.
9. Context and Annotation
Definition:
Charts should provide context through titles, labels, legends, and sources.
Why it matters:
• Helps viewers understand what they are looking at.
• Without context, numbers and visuals may be meaningless.
How to achieve context and annotation:
• Include clear titles, axis labels, and legends.
• Provide units of measurement.
• Add source information when necessary.
Example:
• Title: “Average Rainfall in Kathmandu (mm) – 2025”
• X-axis: Months, Y-axis: Rainfall in mm
Common mistakes:
• Missing axis labels, legends, or units.
• Ambiguous or misleading titles.
10. Storytelling
Definition:
A visualization should tell a story or highlight insights instead of just showing raw data.
Why it matters:
• Helps viewers understand trends, patterns, and decisions.
• Makes data actionable.
How to achieve storytelling:
• Organize charts in logical order.
• Highlight trends, comparisons, or key takeaways.
• Add captions or notes to explain insights.
Example:
• A report starts with total sales trend, then breaks down by region to explain why growth
occurred.
Common mistakes:
• Presenting disconnected charts with no narrative.
• Focusing on aesthetics over insight.
11. Interactivity
Definition:
Interactive visualizations let users explore data dynamically, e.g., filtering, zooming, or drilling
down.
Why it matters:
• Allows deeper exploration of large datasets.
• Improves user engagement and decision-making.
How to achieve interactivity:
• Use dashboards with filters, sliders, and tooltips.
• Allow drill-downs to see details behind summary data.
• Provide hover info to explain points.
Example:
• A dashboard where you can filter sales by product, region, or month to see detailed
trends.
Common mistakes:
• Overcomplicating with too many interactive elements.
• Interactive features without clear purpose.
Gestalt Principles in Data Visualization
Definition:
Gestalt principles are a set of psychological rules describing how humans naturally perceive
visual elements as organized patterns or wholes, rather than as separate parts. Using these in
visualization helps viewers interpret charts quickly and correctly.
1. Proximity
• Objects placed close together are perceived as a group.
• Example: In a scatter plot, clusters of points that are close together suggest a relationship
or group.
2. Similarity
• Items that look similar (color, shape, size) are perceived as part of the same group.
• Example: In a bar chart, all bars representing profits could be green, while losses are red.
3. Continuity
• The eye follows lines or curves naturally. Elements aligned on a line are perceived as
related.
• Example: A line chart showing sales growth is perceived as a continuous trend rather
than disconnected points.
4. Closure
• The brain tends to fill in missing parts to perceive a complete object.
• Example: In a donut chart, even if a section isn’t fully closed, viewers can still perceive
the proportion.
5. Figure-Ground
• The mind separates an object (figure) from its background (ground).
• Example: Highlight a bar in a muted chart; the highlighted bar stands out against the rest.
6. Symmetry & Order
• Symmetrical or orderly visuals are easier to process and appear more organized.
• Example: A well-aligned table or grid of charts looks cleaner and easier to understand.
7. Common Fate (or Movement)
• Elements moving in the same direction or changing together are perceived as related.
• Example: In a line chart with multiple series, lines trending up together indicate a shared
pattern.
. What is Anscombe’s Quartet?
• It is a set of four datasets created by the statistician Francis Anscombe in 1973.
• Each dataset has 11 (x, y) points.
• All four datasets have almost identical statistical properties:
o Same mean of x and y
o Same variance of x and y
o Same correlation between x and y
o Same linear regression line
2. Why it Matters
• Even though the summary statistics are identical, the datasets have very different
distributions and patterns.
• This demonstrates that relying only on numbers can be misleading.
• Visualization is necessary to see trends, outliers, or non-linear relationships.
Key Lessons
1. Never rely solely on summary statistics.
2. Visualization reveals patterns, trends, and anomalies.
3. Outliers and non-linear patterns can drastically affect analysis.
4. Useful for teaching statistics, regression, and the importance of charts.
1. What is Data-Ink Ratio?
• Introduced by Edward Tufte, a pioneer in information design.
• Definition: The proportion of “ink” (or pixels) in a chart that actually represents data,
versus ink used for decoration or non-essential elements.
• Formula (conceptual):
Data-Ink Ratio=Ink used to show dataTotal ink in the chart\text{Data-Ink Ratio} =
\frac{\text{Ink used to show data}}{\text{Total ink in the chart}}Data-
Ink Ratio=Total ink in the chartInk used to show data
• Higher ratio → more of the chart is devoted to actual data, less to decoration.
2. Why It Matters
• Reduces clutter and distraction.
• Makes charts simpler, clearer, and easier to read.
• Focuses the viewer’s attention on important insights rather than decoration.
3. Principles Related to Data-Ink Ratio
1. Maximize data-ink: Every mark should represent data.
2. Minimize non-data ink: Gridlines, borders, 3D effects, and unnecessary colors should
be reduced.
3. Erase redundant data: Avoid repetition of the same information unnecessarily.
4. Examples
Chart Type /
Data-Ink Issue How to Improve
Situation
3D Pie Chart 3D shading adds ink but no extra info Use 2D pie chart
Grid-heavy Bar Chart Thick gridlines distract from bars Lighten or remove grids
Decorative Borders & Borders, textures, background images Remove background, focus on
Patterns add ink but no data bars/points
Labeling every data point when
Repeated Labels Label key points, use axes
unnecessary
5. Example (Simplified)
Bad (Low Data-Ink Ratio):
• Bar chart with thick 3D bars, heavy gridlines, colorful shadows, logos, and unnecessary
text.
Good (High Data-Ink Ratio):
• Simple flat bars, light gridlines, clear axis labels — most ink represents actual data.