Visual Display of Data Brad McMillen Evaluation and Research Department
Visual Display of Data
Brad McMillenEvaluation and Research Department
Things to Consider
Choices:• Text• Tables• Charts/Graphs
What do you want to say?
Who are you saying it to?
What is the most concise way to say it?
Taken from http://www.cdc.gov/nccdphp/dnpa/obesity/trend/maps/index.htm
Example
Text
Between 1996-97 and 2006-07, the number of students enrolled in WCPSS grew by 50%. During
that same time, however, the number taking AP exams increased by 175% and the number of exams taken increased 195%.
Table
Simple, but effective Can be one or two-dimensional (rows and/or
columns) Good for displaying smaller (or larger!) amounts of
data Layout of table can determine how the audience
consumes the data
Bar/Column Chart
Usually used to show changes over time or comparisons among groups/categories
Clustered Column - compares values across categories
Stacked Column - shows the relationship of individual items to the whole
100% Stacked Column - compares the percentage each value contributes to a total across categories
Percent of Students Eligible for Free or Reduced Price Lunch, 2006-07
9.8% 8.4% 7.1%
51.1%
41.7%
31.0%
0.0%
20.0%
40.0%
60.0%
80.0%
100.0%
Elem MS HS
Grade Span
Pe
rce
nt
of
AD
M
Free
Reduced
Line Chart
Shows changes or trends over time Either for a single category or multiple
Horizontal axis usually is a time measurement
Usually with equal intervals of time
Pie Chart
Shows the proportionality across categories as they relate to a whole
Pie - displays the contribution of each value to a total Exploded Pie - displays the contribution of each value to a
total while emphasizing individual values
Percent of Students ProficientGeometry EOC Test, 2006-07
13%
17% 57%
13%
Level ILevel IILevel IIILevel IV
Scatterplot
Shows pairs of values plotted according to X-Y coordinates
• Good for illustrating the relationship between two variables
• Good for displaying the amount of “spread” in the data
50
60
70
80
90
100
80 85 90 95 100
Attendance Rate
Per
cent
Pro
fici
ent
State Math Scores and Students' TV Viewing Habits
AL
AZ
ARCA
CO
CT
DE
DC
FLGA HI
IDIN
IA
KY
LA
ME
MD
MA MI
MN
MS
MO
NENH
NJ
NM
NYNC
ND
OHOK
PA
RI
SCTN
TX
UT
VA
WV
WI
WY
230
240
250
260
270
280
290
0 5 10 15 20 25 30
% Students watching TV 6 hrs+ per day
1992
Sta
te 8
th-g
rade
NA
EP
Mat
h S
core
source: National Center for Educational Statistics, 1994
Histogram
• A chart (usually a simple column chart) that takes a collection of measurements and plots the number of measurements (called the frequency) that fall within each of several intervals
Distribution of EOG Scale Scores
0
10
20
30
40
50
60
239
241
243
245
247
249
251
253
255
257
259
261
263
265
267
269
271
273
275
Principles of Good Graphics
• Clearly labeled
• Not too “busy”
• Readable font
• Quickly draws attention to the desired point
• Reasonable choice of metric and precision
# Students by Ethnicity
0
10000
20000
30000
40000
2003 2004 2005 2006
Asian
African-American
Native American
White
Multiracial
Hispanic/Latino
Report Cards
Taken from http://www.cdc.gov/nccdphp/dnpa/obesity/trend/maps/index.htm
Example
Sources of Deception in Graphics
• Labels not specific
• Scaling of axes is inappropriate
• Choice of chart type “leads” reader to the wrong conclusions
• Ignores important contextual factors
• Too much stuff – obscures the point
Taken from http://mediamatters.org/items/200503220005
Taken from Tufte, E. (1983). The Visual Display of Quantitative Information. Cheshire, CT: Graphics Press.
30
40
50
60
70
80
90
100
WCPSS 75 77 79
NC 74 71 71
USA 49 48 48
2005 2006 2007
WCPSS
USA
NC
EOG Mathematics Proficiency, Grades 3-8
84.589.0 89.5 88.3
64.1
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
2001-02 2002-03 2003-04 2004-05 2005-06
Brad McMillen
Evaluation and Research Department
Wake County Public Schools
(919) 850-1903