Chapter 10: Data DisplayTable, Graphs, Maps, Visualizations
Book:An Introduction to Scientific Research Methods in Geography(Montello & Sutton) 2006GEOG4020-Research MethodsInstructor: Paul C. SuttonUniversity of Denver, Dept. of GeographyPrepared by: Katie WilliamsFebruary 9, 2010
Chapter 10 Overview
Principles of Data DisplayGuidelinesfor Designing DisplaysTablesGraphsMapsNew Trends in Scientific Visualization
2
Learning Objectives
Understand the best use and design alternatives for tables and graphsUnderstand principles of good graphingUnderstand the powers and pitfalls of mapsExplore how computer technologies are applied to display data in innovative and powerful ways
3
Data Display
Depictrather thandescribedata patternsAid in understanding & communicating dataHighlight and clarify relevant data propertiesPurposes of Data Display:ExamineInterpretCommunicate
4
Exploratory Data Analysis
Evaluating how “well-behaved” the data isWhat is the distribution of the data?How homogenous is the data?Does the data fit expected values?Are there extremes (outliers)?Are there impossible values?Submersive, graphical, ad hoc approach
5
Correlation?
Communication of Data
Written and OralArticles, books, papers, talks, interviews, etc.ArchivalViewed more frequently by more peopleRelatively permanent recordsDisplays should be constructed more thoughtfully and thoroughly
6
Data Display
Improper Data Display:Abstract concepts—often better explained with wordsSmall amount of data—in text explanation is sufficientNon-data DisplayCommunicate information other than dataEquipment usedMaterial employed
7
Designing a Display
Goal:Effective CommunicationAccess the complex; not complicate the simpleDepict valid, relevant information clearly, accurately, and unambiguouslyAestheticsAttractiveness draws people’s attentionBut, communication should never be compromised for aesthetics
8
Tables
Table: organized lists, arrays, or matrices of dataOnly minimum use of spatialityGood choice to show data precisely and in detailRound adequatelyTwo depictions:Distribution TablesDescriptive Index Tables
9
Table Types
Distribution TablesFrequency, relative frequency, cumulative frequency, relative cumulative frequencyContingencytables—show relationships between nominal variables or metric variables that can be grouped into discrete classesDescriptive Index TablesCentral tendency, variability, relationship, etc.Organized into rank or class
10
Graphs
Graph: pictorial representation of dataEffective for communicating general rather than precise patterns(especially useful with large datasets)Three Dictums:1.Clearly and sufficiently label the graph and its parts2. Avoid uninformative and content-free marks3. Fill the graph space with data marks
11
Graph Styles
Distribution graphs: depict distribution of variablesValue of variable on x-axis (abscissa)Frequency of occurrence on y-axis (ordinate)Types of Distribution GraphsBar graph—discrete graph styleHistogram—bar graph with quantitative class binsCircle diagram—nominal variable levelLine—continuous variablesCurve-fitting—statistical model fitted to data distribution
12
More Graph Styles
RelationshipGraphs:depictthe form and strength of relationship between pairs of variablesTypes:Scatterplot—plot ofX,Y intersection of two variablesTernary diagram—relationship among three variablesSmall multiples—repeating graph showing change over timeSimulated 3D—graphing data in three dimensions
13
Maps
Maps: graphic displays that depict earth-referenced features and dataQuintessential geographic displayReference Map—Depict earth features accurately and preciselySignificant features are large, stable, & relevantEncoded in coordinate systemThematic Maps—Special-purpose displaysSpatial distribution of thematic variablesLittle earth-surface detail; “map graphs”
14
Map Issues
Any flat rendering of the earth’s surface will result indistortionProjectionsare different methods to flatten the earth while minimizing distortionExamples: Mercator, Sinusoidal (hundreds more)Selectivityis required; one projection will not minimize distortion for the entire surface
15
More Map Issues
Generalization of level of detailSimplification, selection, enhancementMap SymbolsIconic: closely resemble reality (e.g. spatial layout of earth)Abstract: not representative of reality (e.g. contour lines, checkered patterns, words)Feature representationColor, symbols, classes,choroplethregions
16
Principles of Good Mapping
1. Facilitate effective & efficient communication2. Choose relevant and high-quality data3. Show data clearly and truthfully4. Highlight the important & downplay less important5. Focus on the data, not decoration6. Make good choices for map symbols
17
New Trends in Scientific Visualization
Computer driven innovationsThematic mappingComputationally intensive analyticsCapacity for complex and prolific amounts of dataGreat opportunities, but also great challenges
18
Exploratory Data Analysis Techniques
Information visualizationGeo-visualization—visualizing data against geographic backgroundSpatialization—simulated landscapesAnimations—dynamic displays of change over timeAugmented reality—digital displays over actual surroundingsVirtual reality—simulating placesSonification—proposed sound, touch, & smell maps
19
Questions
When does it make most sense to use tables to display data?What are some principles of good graphing and specific design guidelines that derive from them?Selectivity, projection, generalization, and varying symbol abstractness are always involved in mapping. Why are they always involved and why are them potentially misleading to map viewers?
20
0
Embed
Upload