Data visualization: history and principles

Data Visualization

• Key facts are often hidden in data, images clarify (Stephen Few, 2015)
• Data visualization helps analyze complex data, reveal patterns/trends, and improves data-driven decision-making

Historical Evolution

• Early data visualizations used maps
• Claudius Ptolemy (2nd century AD) used coordinate systems to map stars
• René Descartes (17th century) introduced coordinate planes
• William Playfair (18th century): Developed bar charts, line graphs, and pie charts
• Charles Minard (19th century): Created an early dashboard visualizing Napoleon's march to Moscow
• John Snow (19th century): Used maps to identify cholera outbreaks
• Florence Nightingale (19th century): Used visualizations to highlight preventable deaths
• 20th-century advances include:
  • John Tukey: Introduced the box plot and histogram
  • Edward Tufte: Established data visualization as a science (Data-to-Ink Ratio)
  • Stephen Few: Defined best practices and introduced the bullet chart

Learning Objectives

• Craft compelling data-driven stories
• Create audience-focused visualizations
• Develop questions to guide analysis
• Use Tableau to build dashboards
• Quotes from Albert Einstein and Oprah Winfrey emphasize the importance of questioning and audience consideration

Principles of Data Visualization

• Simplify insights
• Compare to highlight differences
• Explore data in different ways
• Be skeptical and check accuracy
• Know your audience
• Make insights easy to understand.
• Help audiences see key differences.
• Question assumptions and check data.
• Tailor stories to decision-makers

Common Pitfalls

• Overcomplicating visuals
• Using pie charts incorrectly
• Not questioning inherited reports
• Avoid visual complexity and errors in graphs
• Address organizational mythology.
• Stephen Few's methods of data visualization

Stephen Few's Core Methods

• Time Series
• Ranking
• Part-to-Whole
• Deviation
• Distribution
• Correlation
• Nominal Comparison
• Geospatial Visualization

Key Takeaways

• Data visualization focuses on storytelling, not just charts
• History impacts modern visualization techniques
• Tableau is a primary practical learning tool
• Effective visualizations are simple, insightful, and audience-focused
• The team final project carries a significant grade portion

Learning Objectives

• Notes from the second class.
• To learn data visualization

Story Presentation

• Engaging story presentation
• Audience-focused data visualization
• Exceptional question design

First Data Displays

• Babylonian Clay Tablet (ca. 1400 BCE) depicts the city of Nippur
• Babylonian Map of the World (ca. 500 BCE) reflects Babylonian knowledge

Early Data Visualization

• Claudius Ptolemy (150 AD) wrote "Almagest" to record astronomy, the sun, moon, and planet movements

René Descartes

• René Descartes (1596-1650) invented a 2D coordinate system

William Playfair

• William Playfair (1759-1823) improved statistical data visualization
• Created the time-series line graph, bar chart, and pie chart

Charles Minard

• Charles Minard's (1781-1870) map of Napoleon's march on Moscow is a visualization of troop depletion.
• It uses line width to show army size changes, showing troop losses through maps, time series, and temperature data

John Snow & Nightingale

• John Snow (1813-1858) used a map of London cholera outbreak in 1854 to identify the source
• Florence Nightingale (1820-1910) used coxcomb diagrams to show preventable soldier deaths in the Crimean War, improving field hospital conditions

Joseph Priestley

• Joseph Priestley's "Chart of Biography" visualized 2000+ historical figures
• The timeline combines historical civilizations and sets the stage for modern timeline visualizations

John Tukey & Edward Tufte

• John Tukey (1915-2000) published Exploratory Data Analysis in 1977 and introduced box plots, histograms, Pareto charts, scatter plots, etc
• Edward Tufte is known as the pioneer in data visualization who published "The Visual Display of Quantitative Information" in 1983

Data Visualization Principles

• Showing the data
• Focuses on substance rather than design
• Avoids distorting data
• Presents many data in a small space
• Makes large datasets coherent
• Encourages comparing data
• Reveals data at multiple levels (overview to detail)
• Integrates statistical and verbal descriptions

Stephen Few

• Stephen Few modern data visualization best practices include the information dashboard design and bullet charts

Quantitative Data

• There are eight types of displaying quantitative data
• Time-Series
• Ranking
• Part-to-Whole
• Deviation
• Frequency Distribution
• Correlation
• Nominal Comparison
• Geographic/Geospatial

Data Analysis

• There are eight major principles of data analysis
• Simplify
• Compare
• Attend
• Explore
• View Diversely
• Ask Why
• Be Skeptical
• Respond and Share

Visual Perception Importance

• The human visual system is capable of identifying data patterns
• Data visualization exploits visual cognition, easing understanding and analysis

Visual Perception Process

• Light enters the eye, focusing on the retina
• Rods help with night vision
• Cones support color under bright conditions
• Retinal cells convert light to electrical signals via the optic nerve
• The brain processes information to create a 3D image and interprets visual characteristics

Visual Perception We Can Leverage

• Aspects used of Memory and Retention
• Sensory Memory- deals with initial processing of stimulus
• Working Memory- processes information and stores it briefly
• Long-Term Memory- stores semantic, recognition, and cognitive processes, can go in for a lifetime

Goals

• Capture attention through visual perception (sensory memory)
• Activate short-term memory (working memory)
• Transfer to Long-Term Memory

Pre-Attentive Attributes

• Features processed automatically without conscious effort
• Attributes aid data detection/discrimination, important in design
• Main traits include color, orientation, size, motion, and spatial grouping (Gestalt Principles)

Gestalt Principles of Perception

• Principles emphasize how the brain organizes info to create meaningful patterns
• Proximity: Elements placed close together creates a group
• Similarity: Elements with similar shapes/colors creates a group
• Enclosure: Borders highlight single elements
• Continuity: Eyes follow continuous data paths
• Connection: Lines/shapes connect to be a single group
• Closure: Brain fills incomplete shapes
• Symmetry and Order: Eyes prefer balanced structures
• Figure-Ground: Visually identifies the foreground from the background

Encoding Data

• Pre-Attentive Quiz tests attentiveness, let students experience various data encoding
• Encoding with color, size, and borders making recognition easy
• The goal is to help audiences quickly and precisely comprehend information
• Optimize data using visual perception simplifies to see patterns, trends, and outliers

Tables

• Needed when exact values and individual numbers are looked for
• Compares specific numbers and accurate data is a must have
• Used for financial and scientific data
• Tables can use many measurements
• Gives summaries and detailed data

Graphs

• Used for trends
• Shows relationships in complete data

Visual Marks

• Points
• Lines
• Bars
• Boxes
• 2-D Shapes
• Color Intensity

Charts

• Must start at zero
• Start at zero to prevent misinterpretation

Foundations of Storytelling

• Important to do
• Basic Chart Types
• Pie Charts
• Bar Charts
• Histograms
• Bullet Charts
• Stacked Bar Charts
• Line Charts
• Scatterplots
• Tables
• Cross tabs
• Highlight tables
• Additional Visualizations
• Area charts
• Bubble charts
• Gantt charts
• Heat maps
• Treemaps
• Charts
• Candlestick
• Charts
• Mapping and Geospatial
• Infographics

Common Chart Types

• Each chart has a specific use
• Pie Charts- shows proportions and data distribution
• Bar Charts- Compares many datasets
• Histograms- Shows the spread of data, can pinpoint outliers

Bullet Charts

• Used to display data, normally dashboard gauges
• Stacked Bar Charts compare internal sets of data
• Line Charts display highs and lows of data

Relationships

• Scatter Plots shows the relation between each bit of data
• Treemaps organizes data from top to bottom
• Maps shows geological data
• Box and Whisker shows the layout of data

Charts with Uses

• Bubble Charts- Shows data for project management that shows start and end in a project
• Candlestick Charts- Shows how best to breakdown the stock market

Data Filters

• Filters helps break down data
• Easy to highlight key trends
• Support advanced analysis
• Filter types
• Extract Filters
• Datasource Filters
• Context Filters
• Dimension Filters
• Measure Filters
• Set Filters
• Conditional Filters
• Top Filters

Visualizing Relationships

• To know how different visual methods show data
• How choose what chart to support the story

Form Elements

• Color and form are important
• Color Theory includes tone (Hue), strength (Intensity), saturation, or brightness
• Color choice brings particular meanings

Presentation

• Colors help for specificreasons
• Sequential Color
• Diverging Color
• Categorical Color
• Highlighting Color
• Alerting Color
• Lowers data information by changing up colors
• Balances data

Color Perception & Accessibility

• Color perception matters
• Normal vision processes colors well
• Some have a hard time with seeing colors properly

Optimized Viewing

• Optimizing for color blindness is essential

Visualizing Relationships

• Minard's famous map
• Comparisons on data scale
• Shows casualty, mechanism, structure, and explanation
• Multivariate, analysis, and evidence

Multivariate

• There are 6 pieces to multivariate
• Size of army
• Dimension
• Direction
• Temperature
• Date

Documentation

• Documentation helps the reader
• Needed pieces for understanding

Measure

• Used in class 6
• Measure is used with names and vales
• Tableaus help with values

Totals

• Easy break downs of tableaus

Sum

• Sets and data
• Sorts specific pieces of data
• Made inside and outside

Manual Production

• Done to combine each different set
• Used for combining 2 or more sets

Table Calculations

• Analysis is done on a quick and basic level with the tables
• Scopes are useful for direction of data
• Helps with many Tableau charts

LOD

• Has different types
• Fixed types

Storytelling

• The key elements of storytellings, use different elements used in the beginning, middle and end
• Class 7 is where this is from

Attention

• It needs to capture the audience to captivate it
• Has to determine a purpose

Interactive data

• Very useful to to use with other variables

Data Stories

• Class 4.7
• It needs to have the 7 forms which is:
• Change over time
• Zoom Out
• Drill down
• Contrast
• Intersections
• Factors
• Outliers

Visualizations

• Its a way to limit pitfalls in the data
• It's a way to limit the type of charts used
• Chart examples that should not occur:
• Worst charts
• Charts with a wrong axis
• Charts with an over abundance of clutter
• Used to help organize the data