1. What is Data Visualization?
Definition:
Data visualization (or “DataViz”) is the art and science of communicating data
visually—using images, charts, and graphics that are readable and recognizable.
Purpose:
To make invisible data visible — turning raw numbers into insights people can
understand.
Core idea:
It blends art (design) and science (data analysis) — this balance is called creativity.
🎨 2. Why Data Visualization Matters
The most effective visualizations use storytelling.
Stories make data:
1. Memorable – People remember visuals better than numbers.
2. Relatable – Creates emotional connection.
3. Actionable – Inspires decision-making and action.
📊 3. Two Main Types of Data Visualization
Type Purpose Example
To explore data and find patterns, Analyzing datasets like the
Exploratory trends, or relationships. Often usedAnscombe’s Quartet to see how visuals
when problems are not yet clear. reveal differences not visible in tables.
Dr. John Snow’s 1854 cholera map
To explain findings and communicate
Explanatory showing deaths linked to a water
insights clearly.
pump.
➡️Exploratory = Discover
➡️Explanatory = Explain
🧩 4. Visualization Design Principles
a. Gestalt Principles (Psychology of Visual Perception)
These explain how people naturally group visual elements to make sense of images.
Principle Meaning Example
Things close together are seen as a Dots near each other look
Proximity
group. grouped.
Principle Meaning Example
Same-colored bars feel like one
Similarity Similar shapes or colors appear related.
category.
Boxes around data create visual
Enclosure Boundaries or shaded areas group items.
groups.
Connected elements are perceived as
Connection Lines linking points.
related.
Symmetry Balanced visuals are seen as one whole. Mirror shapes form pairs.
The eye follows continuous lines or
Continuity Line graphs.
paths.
We mentally fill gaps to see complete Seeing a full circle from broken
Closure
shapes. lines.
Figure & Focus shifts between background and
Seeing either a vase or two faces.
Ground main figure.
b. Edward Tufte’s Design Principles
A pioneer in data visualization (Yale University).
Key ideas:
1. Show the data clearly.
2. Maximize data-to-ink ratio – only include what adds value.
3. Remove “chartjunk” – avoid unnecessary colors, effects, borders, or
shadows.
4. Clarity, precision, and efficiency are the goals.
o ➡️Less is more.
5. Pre-attentive Attributes
These are visual features our brain processes instantly—without conscious effort.
a. Emphasis (What stands out)
Use form, position, motion, or color to draw attention.
o Form: length, width, shape, size, curvature, blur, enclosure.
o Position: where you place data on the chart.
o Motion: direction or movement (trend arrows).
o Color: hue or intensity to highlight focus areas.
b. Quantity
Some visual tools show quantity more accurately than others:
o Most accurate: position.
o Least accurate: color saturation or density.
c. Color
Used to emphasize, categorize, or convey meaning.
Best practices:
1. Avoid overload – too many colors confuse.
2. Be consistent – use the same color for the same category.
3. Convey meaning – e.g. green = good, red = bad (but be mindful of color blindness
and cultural differences).
4. Use brand colors when presenting for a company.
5. Darker colors = larger quantities (in scales or heatmaps).
📈 6. Choosing the Right Chart
Different charts serve different purposes:
Purpose Example Charts
Compare values Bar chart, column chart
Show trends over time Line chart, area chart
Show proportions Pie chart, stacked bar
Show relationships Scatter plot, bubble chart
Show distribution Histogram, box plot
Show hierarchy Tree map, sunburst
Show location Map
Show flow/process Sankey, flowchart
👉 Use tools like [Link] to find the right chart for your purpose.
📱 7. Visualization Viewing Patterns
F-pattern: used when reading vertical layouts (e.g. phones).
Z-pattern: used when reading horizontal layouts (e.g. computer screens).
Design accordingly for where your visualization will be viewed.
📖 8. Storytelling in Data Visualization
Narrative Framework (from Stanford’s Interactive Data Lab):
1. Visual Design – How you show your story (colors, layout, charts).
2. Messaging – How you tell your story (titles, captions, summaries).
3. Interactivity – How your audience explores your story (filters, sliders, dashboards).
Structure of a good data story:
Beginning: Background and context.
Middle: Problem or rising action (conflict).
Climax: Key insight or discovery.
End: Resolution or conclusion.
✍️9. Author-Driven vs Reader-Driven Approach
Approach Control Messaging Interactivity Best for
Author- Author controls High (clear Executives,
Low
Driven sequence insights) presentations
Reader- Reader explores
Low (exploratory) High Analysts, dashboards
Driven freely
Exploratory visualizations → Reader-driven
Explanatory visualizations → Author-driven
👥 10. Understanding Your Audience
Audience Role Recommended Approach
Executives (C-level) Decision-makers Author-driven, concise, outcome-focused
Departmental Managers Middle-level Balanced approach, guided interactivity
Reader-driven, full interactivity, deep
Analysts Data experts
exploration