visualization - carnegie mellon university

39
Visualization Scott Donaldson | Adv. User Interface Software | CMU HCII 2017

Upload: others

Post on 23-Oct-2021

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Visualization - Carnegie Mellon University

VisualizationScott Donaldson | Adv. User Interface Software | CMU HCII 2017

Page 2: Visualization - Carnegie Mellon University

1. What is it?

2. How did we get here?

3. How do we do it today?

Page 3: Visualization - Carnegie Mellon University

WHAT IS IT?Data Visualization

Page 4: Visualization - Carnegie Mellon University

DATA VISUALIZATION

Page 5: Visualization - Carnegie Mellon University

DATA VISUALIZATIONStatistics

MathematicsScientific Research

Census DataPolitical/Economic

Page 6: Visualization - Carnegie Mellon University

DATA VISUALIZATIONGraphical Representation

Visual Variables(A Much Newer Notion)

Page 7: Visualization - Carnegie Mellon University

HOW DID WE GET HERE?Data Visualization

Page 8: Visualization - Carnegie Mellon University

Early Visualization Methods

Symbolic Representation

Lists & Tables

2-Dimensional Graphing

Cartography

Page 9: Visualization - Carnegie Mellon University

Symbolic RepresentationFirst step toward creating data visualization is defining visual variables (numbers, letters, icons) that communicate quantities, qualities

Ex. Hieroglyphics

Page 10: Visualization - Carnegie Mellon University

Information displayed according to a single dimension (alphabetical order, index, size, location)

Ex. Domesday Book (1086): “Great Survey” of England

Lists & Tables

Page 11: Visualization - Carnegie Mellon University

New dimension allows greater freedom in representing data — independent axis corresponds to list, data measured by dependent axis

Ex. Cartesian coordinate system (René Descartes, 1637)

2-Dimensional Graphing

Page 12: Visualization - Carnegie Mellon University

Adding ‘metadata’ to maps like Cartesian coordinates — creates higher dimensions to display more complex information

Ex. Edmund Halley’s map of magnetic variations (1702)

Cartography

Page 13: Visualization - Carnegie Mellon University

Development of Foundational Visualization Techniques

Bar Chart

Pie Chart

Timeline

More Complex Representations

Page 14: Visualization - Carnegie Mellon University

Pie Chart, Bar Chart, TimelineAlmost all basic formats of visualization invented by William Playfair (1759-1823), Scottish engineer, political economist

Page 15: Visualization - Carnegie Mellon University
Page 16: Visualization - Carnegie Mellon University

Building on Foundations

After Playfair, others began combining types of graphics into more complex visualizations

Ex. Charles J. Minard (1781-1870), French civil engineer

Page 17: Visualization - Carnegie Mellon University
Page 18: Visualization - Carnegie Mellon University
Page 19: Visualization - Carnegie Mellon University
Page 20: Visualization - Carnegie Mellon University

Modern Developments

Computers and Graphical User Interface

Visualization as Autonomous

Building on This History

Page 21: Visualization - Carnegie Mellon University

Pre-1960s, computers but no graphical user interfaces

Ivan Sutherland, Sketchpad, 1962-63

Demonstration video

Computation and Visualization

Page 22: Visualization - Carnegie Mellon University

GUIs introduced in the late ‘60s and ‘70s but did not become widespread until adoption of PCs in workplaces in the ‘80s

Ex. Xerox Star (1981)

Developing the Graphical User Interface

Page 23: Visualization - Carnegie Mellon University

Digital visualization did not invent a language of representation; inherited it from Playfair, Minard, Nightingale

Ex. Lotus 1-2-3 Graph Maker (1983)

Using Computers toVisualize Data

Page 24: Visualization - Carnegie Mellon University

Simultaneous to GUI development, new theories of visualization apart from the data it represented were being formulated in the academy

Ex. Jacques Bertin, Semiologie Graphique (1967); described elements of visualization as parameters, variables (language of computation)

Visualization asAutonomous

Page 25: Visualization - Carnegie Mellon University

Edward Tufte’s The Visual Display of Quantitative Information (1983) identified the lineage of Playfair, Minard, Nightingale, and others

Made visualization accessible to a broader audience than statisticians/engineers, to include businesses and the general public

Creating the Historyof Visualization

Page 26: Visualization - Carnegie Mellon University

HOW DO WE DO IT TODAY?

Data Visualization

Page 27: Visualization - Carnegie Mellon University

Contemporary Software for Data Visualization

Proprietary, Business-Oriented

AnyChart (2003)

FusionCharts (2003)

Tableau (2005)

Open Source, General Purpose

R Language (1993)

Processing (2001)

Protovis (2009) → D3.js (2011)

Page 28: Visualization - Carnegie Mellon University

Originally Flash, now HTML5 + JavaScript

JavaScript library + API (no GUI) (Documentation)

Charts: Area, Bar, Box, Bubble, Bullet, Column, Doughnut, Error, Funnel, Heat Map, Japanese Candlestick, Jump Line, Line, Marker, OHLC, Pareto, Pie, Polar Plot, Pyramid, Radar, Scatter Plot, Sparkline, Spline, Spline Area, TreeMap, 3D Bar, 3D Doughnut, 3D Pie, and more

AnyChart

Page 29: Visualization - Carnegie Mellon University

AnyChart Example

Page 30: Visualization - Carnegie Mellon University

Desktop app/web app ($, but free version with limited features)

Drag and drop GUI, some customization but difficult to adjust fine-grain details

Charts: Bar, Table, Line, Scatter Plot, Heat Map, Histogram, Gantt, Pie, Treemap, Box Plot, Packed Bubble, Map (many more not listed on website)

Tableau

Page 31: Visualization - Carnegie Mellon University

Tableau Example

1. Connect to data source2. Drag the Order Date dimension

to Columns3. Drag the Sales measure to

Rows4. On the Marks card, select Bar

from the drop-down list5. Drag the Ship Mode dimension

to Color on the Marks card6. Drag the Region dimension to

Rows, and drop it to the left of Sales to produce multiple axes for sales by region

Page 32: Visualization - Carnegie Mellon University

JavaScript library + extensions for jQuery, AngularJS, PHP, Asp.NET, JSP, Django, React, Ruby on Rails

Declarative, not imperative: Write configuration markup (JSON, XML, or HTML + JS)

Powers Federal IT Dashboard, Google Docs chart widget

FusionCharts

President Obama is a fan!

Page 33: Visualization - Carnegie Mellon University

FusionCharts Example

Page 34: Visualization - Carnegie Mellon University

Open source language for statistics, analysis, and visualization

~8,000 packages, many specifically visualization-oriented (ggplot2)

Operates efficiently at scale (millions of data points), favored by data analysts both for computing power and visualization

R Language

Page 35: Visualization - Carnegie Mellon University

R Language Example

Page 36: Visualization - Carnegie Mellon University

Java-based IDE for “learning how to code within the context of the visual arts”

Libraries for JavaScript, Python

Plugins/libraries for audio, video, 3d visualization, physical devices (Arduino, Raspberry Pi)

Processing

Page 37: Visualization - Carnegie Mellon University

Processing Example

Sorry, no bar chart to show!

Moral of the story: lots of coding “by hand” to do simple things

Page 38: Visualization - Carnegie Mellon University

D3.jsAcronym for “data-driven documents.” JavaScript library for “manipulating documents based on data.”

Steep learning curve, but extremely flexible/extensive

Gained popularity through New York Times interactive projects

Page 39: Visualization - Carnegie Mellon University

D3.js Example