1 Lesson 3: Presenting Data Graphically

1.1 Types of graphs

Once data is organized and arranged, it can be presented. Graphic representa-
tions of data are called graphs, plots or charts. There are an untold number of
graph and chart types, and a myriad of names for these types. We will discuss
a few of the more popular or common types, and we will use the names for them
in many spreadsheets.

1.1.1 Scatter Plots

The simplest type to start with are scatter plots. These are very popular in
mathematics applications, and are a good choice when your data comes as pairs
of numbers. (Each datum is a pair of numbers.) In a real application, the more
data points drawn that can be plotted on a computer the better; in a hand
drawn classroom exercise, the fewer the better.
First a middle school example:
Bags of jelly beans come in various sizes, and cost various amounts of money.
The cart below gives the costs of bags of various sizes.

Number of Beans 100 200 250 400 500 750

Cost $1.50 $3.00 $3.75 $6.00 $7.50 $11.25

First we must make a note of the organization of the data:

The data is organized as pairs of numbers written in columns, the …rst

column only being labels.
All the entries are numbers.
The data is arranged in increasing order along each row.
There is a functional relationship in both directions of the pairs.

A scatter plot of this data might look like

Jelly Beans
$12.00

$10.00

$8.00
Cost

$6.00
Cost
$4.00

$2.00

$0.00
0 100 200 300 400 500 600 700 800
Number of Beans

1
This has the cost on the vertical axis and the number on the horizontal axis.
There is another choice which switches these positions.

Jelly Beans
$12.00

Number of Beans $10.00

$8.00

$6.00

$4.00

$2.00

$0.00
$0.00 $2.00 $4.00 $6.00 $8.00 $10.00 $12.00
Cost

Notice, to get this second plot, you would say that we have rearranged the
data. We might have rearranged the data by switching the two rows in our data
table to
Cost $1.50 $3.00 $3.75 $6.00 $7.50 $11.25
Number of Beans 100 200 250 400 500 750

Even if we did not actually do it on paper, it should be considered a rearrange-

ment.
What is happening here is that we are choosing between the two functional
relationships we have available. The cost of a bag of jelly beans is in a functional
relationship with the cost of the bag. The "xvariable" is typically drawn on the
horizontal axis, so we might say that the …rst graph illustrates this functional
relationship. If we look at the second graph, the cost is on the horizontal axis.
We can interpret this as illustrating the functional relationship the number of
jelly beans has with the cost. All we are doing is giving a mathematical context
to the only slightly di¤erent graphical representations of the same thing. Mostly
the distinction is just your point of view, but mathematicians do have a their
own preference when it comes to functions.

What if there is no functional relation in a table? Then a scatter plot is

still a good choice. Consider the 2008 Philadelphia Eagles football team again.
There are 80 players on the team mostly, but not all, big men. We can collect
a table of the heights and weights of all these players. Without looking at the
table, we know how it might be organized:

The data is organized as columns of (First Name, Second Name, Height,

Weight)..
The …rst two entries in a datum are names, the second two entries are
numbers.
The data is arranged alphabetical order by last name.

2
 There is a functional relationship in both directions of the pairs.
The players’heights are in a functional relationship with the full names.
The players’weights are in a functional relationship with the full names.
The players’heights are not in a functional relationship with the weights.

The players’ weights are not in a functional relationship with the full
heights.

These last four observations are worth a look at. No two players have the
same …rst and last names, so a player’s full name uniquely identify him and thus
both his height and weight. That is what we need for the …rst two functional
relationships with the player’s full name. Without actually looking at the data,
we are only guessing that players’ heights are not in a functional relationship
with the weights. But them it seems rather likely that in a list of 80 people,
two might have the same weight but have di¤erent heights. One example of
that is enough to eliminate a functional relationship. The last observation is
on more solid ground. We already saw that Andrews and Baskett are the same
height, but di¤erent weights. So we know for certain that the players’weights
are not in a functional relationship with their heights.
If we are interested in the player’s sizes, and not the players themselves, we
can rearrange the data by leaving out the players’names. (Remember dropping
data is just a rearrangement.) Now this leaves a data set made up of numbers,
and better still the numbers are in pairs. Unfortunately, there are no functional
relationships in these number pairs. A scatter plot still works. In fact, the
good side of not having a functional relationship in pairs of numbers is that
there is basically only one type of chart that we can use to display that data: a

3
scatter plot.

Philadelphia Eagles
400

350
300

250
Weight

200
Weight
150

100
50

0
66 68 70 72 74 76 78 80 82
Height

Each diamond represents a player; its place along the horizontal axis is the
players height; it place along the vertical axis represents his weight. We can see
that there are a number of players who are 6’1”tall, and their weights vary from
about 170 pounds to about 350. So this there is not a functional relationship
in this data, but the scatter chart gives a very clear picture of how large these
players are. The scatter plot does not need a functional relationship is provide
good visual information. (Notice how clearly the scatter plot shows the curious
fact that no player is exactly 6 foot tall.)
In summary, scatter graphs as good for data made up of pairs of numbers.
They work whether or not there is a functional relationship.

4
1.1.2 Line graphs
The example of the costs of jelly beans led us to the scatter plot

Jelly Beans
$12.00

$10.00
Number of Beans

$8.00

$6.00

$4.00

$2.00

$0.00
$0.00 $2.00 $4.00 $6.00 $8.00 $10.00 $12.00
Cost

There is such a strong pattern in this data, that it almost jumps o¤ the graph.
These points line up! If we were to accentuate this by drawing a line between
the points, we would have the line graph

Jelly Beans
$12.00
Number of Beans

$10.00

$8.00

$6.00 Cost
$4.00

$2.00

$0.00
0 100 200 300 400 500 600 700 800
Cost

This line graph is possible because the number of jelly beans is in a func-
tional relationship with the cost. For a line graph to work, there must be a
functional relationship. The data of Eagles football players’heights and weights
has nothing resembling a functional relationship, and there is no way of con-
necting the points in its scatter plot into a line graph. Line graphs are only
possible when there is a functional relationship in pairs of numbers in the data.
When the "input" data in the functional relationship has a natural order, line
graphs are often a good choice.
There only has to be a functional relationship in one direction for a line
chart to work. In the last lesson we considered the grades on three tests earned

5
by a group of students.

Name Test 1 Test 2 Test 3

April 55 71 64
Barry 63 67 63
Cindy 88 90 91
David 97 92 87
Eileen 58 55 75
Frank 90 89 96
Gena 88 100 85
Harry 71 70 71
Ivy 65 75 85
Jacob 77 70 65
Keri 75 88 85
Larry 88 92 92
Mary 95 95 100
Norm 86 82 80

At …rst this data does not seem to …t in the pattern of either a scatter plot
or a line plot. However, if we look at one datum from this, we have a di¤erent
situation.
Name Test 1 Test 2 Test 3
Mary 95 95 100
Here the data has the form

Each datum is a pair of the form (Test #, Score)

The score is in a functional relationship with the test number.
The test number is not in a functional relationship with the score.

Here is a set of data that can be illustrated using a line chart.

Mary
100

99

98

97
Grade

96

95

94

93

92
Test 1 Test 2 Test 3

Notice that we placed the tests along the horizontal axis. Typically the "in-
put" item in a functional relationship is set along the horizontal axis, and the
"output" item is placed on the vertical.

6
 Now Mary is a good student, but this graph makes it look she improved
quite a bit on the last test. However, she was an A student throughout, and
this graph may not give the correct impression. We can change the scale of
grades to correct this impression, although the result may be just as bad.

Mary
100
90

80

70

60
Grade 50

40

30

20
10

0
Test 1 Test 2 Test 3

For now, we will not worry that much about adjusting a chart or graph visually,
but rather concentrate on simply selecting the best type of chart. Later we will
revisit the idea of formatting a chart for visual e¤ect.
Still the scale from 0 to 100 might work better for other students, and cer-
tainly if we want to compare several students. Consider the girls:

Name Test 1 Test 2 Test 3

April 55 71 64
Cindy 88 90 91
Eileen 58 55 75
Gena 88 100 85
Ivy 65 75 85
Keri 75 88 85
Mary 95 95 100

7
 We can give each girl her own line graph, and place them all in one chart.

Test Scores
100

90

80
April
70 Cindy
60 Eileen
50 Gena
Ivy
40
Keri
30
Mary
20

10

0
Test 1 Test 2 Test 3

For another example, consider the Department of Education program called

Math and Science Partnerships. The Department distributes funds to the states
for speci…c projects, and the funds received by Arizona over since 2001 are given
in the table:
2001 2002 2003 2004 2005 2006 2007

A riz o n a $ 6 ,7 5 9 ,0 1 3 $ 1 0 ,1 1 4 ,3 4 6 $ 9 ,6 5 5 ,0 5 4 $ 1 2 ,2 0 2 ,5 1 9 $ 9 ,2 7 8 ,8 9 9 $ 5 ,2 9 1 ,6 9 7 $ 5 ,2 9 0 ,4 6 4

The data is organized as pairs written in columns; and so each row is a

datum.
The data can be thought of as made up of pairs of numbers.
The money is in a functional relationship with the year.
The data is arranged in increasing order of years.

Since we have pairs of numbers, we could use a scatter plot. But there is a
functional relationship, so we could just as well use a line plot. Since the years

8
have a natural order, a line plot is a good choice:

Arizona MSP Funds

14000000

12000000

10000000

8000000
Amount

6000000

4000000

2000000

0
2001 2002 2003 2004 2005 2006 2007

Notice, the only lines in this graph are the ones we added. All the line graph
does is connect points in the scatter plot with straight line segments. In many
cases, a line graph is a visual presentation of the data more than a mathematical
representation.
It is possible, and sometimes useful, to smooth out the plot by curving the
lines through the points in the scatter plot. The results is still just a variation
of the line graph. There are times when shading in the area below the lines
in the line graph is a good idea for visual e¤ect, but mathematically such an
"area" chart is only a variation of a line graph. Line graphs are very useful, and
there are any number of major and minor variations of their basic idea that can
come in handy.

1.1.3 Column charts and bar charts

Consider the Arizona MSP data a second time.
2001 2002 2003 2004 2005 2006 2007

A riz o n a $ 6 ,7 5 9 ,0 1 3 $ 1 0 ,1 1 4 ,3 4 6 $ 9 ,6 5 5 ,0 5 4 $ 1 2 ,2 0 2 ,5 1 9 $ 9 ,2 7 8 ,8 9 9 $ 5 ,2 9 1 ,6 9 7 $ 5 ,2 9 0 ,4 6 4

The amount of money is in a functional relationship with the year. The lines
between points in the Arizona MSP funding chart give the impression that the
funds arrived uniformly during the year. This is not necessarily wrong, but it
might not be the impression we want to give. A simple scatter plot looks too

9
sparse, and may be hard to read.

Arizona MSP Funds

$12,000,000.00

$10,000,000.00

$8,000,000.00

$6,000,000.00

$4,000,000.00

$2,000,000.00

$0.00
2001 2002 2003 2004 2005 2006 2007

But because there is a functional relationship, and we can use a column chart.

Arizona MSP Funds

$12,000,000.00

$10,000,000.00

$8,000,000.00

$6,000,000.00

$4,000,000.00

$2,000,000.00

$0.00
2001 2002 2003 2004 2005 2006 2007

The nice thing about column charts is that the "input" data in the functional
relationship does not need to be numbers. You can use column charts when the
data along the horizontal axis are numbers, names, words, or anything. These
data need a functional relationship with numbers, but the results can be quite
nice.
There is no rule that the columns need to be simple; all sorts of shapes can
be used. Three dimensional bars work

Arizona MSP Funds

$10,000,000.00
$9,000,000.00
$8,000,000.00
$7,000,000.00
$6,000,000.00
$5,000,000.00
$4,000,000.00
$3,000,000.00
$2,000,000.00
$1,000,000.00
$0.00
2001 2002 2003 2004 2005 2006 2007

10
We could just a well use cylinders, pyramids, cones, dollar signs, people in cap
and gown, or basically anything we can draw to scale. We can draw it sideways
and turn it into a bar chart:

Arizona MSP Funds

2007

2006

2005

2004

2003

2002

2001

$0.00 $2,000,000.00 $4,000,000.00 $6,000,000.00 $8,000,000.00 $10,000,000.00 $12,000,000.00

Column and Bar charts work very well when there are only a few data items
to illustrate, and even more so when the numbers represent a quantity of some
sort. They are less useful when one numerical value is signi…cantly larger than
all the others. It is also very important to note the scale used when reading
one of these charts. The scale used in a column chart may be chosen to visually
exaggerate a di¤erence in the data. This is not necessarily bad, but it does
require the reader to be alert. Any column or bar chart that does not have
clear numerical labels in its scale should be viewed with scepticism

1.1.4 Pie Charts

Pie charts are a great choice for displaying percentages. Take, for example, the
MSP funding again:

2001 2002 2003 2004 2005 2006 2007

A riz o n a $ 6 ,7 5 9 ,0 1 3 $ 1 0 ,1 1 4 ,3 4 6 $ 9 ,6 5 5 ,0 5 4 $ 1 2 ,2 0 2 ,5 1 9 $ 9 ,2 7 8 ,8 9 9 $ 5 ,2 9 1 ,6 9 7 $ 5 ,2 9 0 ,4 6 4

Note again: The data is organized as pairs written in columns; and so each row
is a datum.

The data can be thought of as made up of pairs of numbers.

The money is in a functional relationship with the year.
The data is arranged in increasing order of years.

Before we build a pie chart, we need to rearrange the data. First, we add a
total column
2001 2002 2003 2004 2005 2006 2007 To ta l

A riz o n a $ 6 ,7 5 9 ,0 1 3 $ 1 0 ,1 1 4 ,3 4 6 $ 9 ,6 5 5 ,0 5 4 $ 1 2 ,2 0 2 ,5 1 9 $ 9 ,2 7 8 ,8 9 9 $ 5 ,2 9 1 ,6 9 7 $ 5 ,2 9 0 ,4 6 4 $ 5 8 ,5 9 1 ,9 9 2

11
A pie chart needs a functional relationship to work, but more important, the
output of the function needs to be a number. In this example, the amount
of funds is in a functional relationship with the year. Each year has its own
funding amount, and that will be represented with its own slice of a pie. The
size of the slice is proportional to the amount of money received in that year.
To draw the pie chart, we need to rearrange our data to include percentages:

2001 2002 2003 2004 2005 2006 2007 To ta l

A riz o n a $ 6 ,7 5 9 ,0 1 3 $ 1 0 ,1 1 4 ,3 4 6 $ 9 ,6 5 5 ,0 5 4 $ 1 2 ,2 0 2 ,5 1 9 $ 9 ,2 7 8 ,8 9 9 $ 5 ,2 9 1 ,6 9 7 $ 5 ,2 9 0 ,4 6 4 $ 5 8 ,5 9 1 ,9 9 2

Percentage 12% 17% 16% 21% 16% 9% 9% 100%

In a pie chart, the pie is made up entirely by the slices. In our case, each slice
represents the funding in one year. The area of each slice is the same percentage
of the whole pie as the percentage of the amount obtained in that year. Now
the areas in pie slices are proportional to the degrees in their angles. Since a
whole pie comes from an angle of 360 we rearrange the table one more time to
give angles of the corresponding percentage:

2001 2002 2003 2004 2005 2006 2007 To ta l

A riz o n a $ 6 ,7 5 9 ,0 1 3 $ 1 0 ,1 1 4 ,3 4 6 $ 9 ,6 5 5 ,0 5 4 $ 1 2 ,2 0 2 ,5 1 9 $ 9 ,2 7 8 ,8 9 9 $ 5 ,2 9 1 ,6 9 7 $ 5 ,2 9 0 ,4 6 4 $ 5 8 ,5 9 1 ,9 9 2

Percentage 12% 17% 16% 21% 16% 9% 9% 100%

A n g le 42 62 59 75 57 33 33 360

After this, it is just a matter of how you like to draw it. A protractor helps,
but not as much as a computer program:

Arizona MSP Funds

2007 8%
2001 14%
2006 8%

2002 17% 2005 14%

2003 19% 2004 20%

There are lots of ways to draw and label the same data:

12
 Arizona MSP Funds
2007 $3,908,204.00
2001 $7,016,251.00
2006 $4,055,169.00

2002 $8,794,248.00 2005 $7,260,015.00

2003 $9,690,136.00 2004 $9,868,971.00

Or if you want to split it up a bit,

Arizona MSP Funds

2007 8%
2001 14%
2006 8%

2002 17% 2005 14%

2003 19% 2004 20%

We can easily see that the largest percentage of funds came in 2004, and the
least in 2006.
Now a pie chart does not need to be a pie, it can be a rectangle:

13
1.1.5 Stem and Leaf Plots; Frequency Graphs and Histograms
There are many other types of graphs than scatter plots, line graphs, column
and row charts, and pie charts. There are also various versions of each of
these types. There are plenty of choices for representing data graphically. We
do not have time to do more than outline the basic types above. There is,
however, one last general type of graphic associated with data. we will consider
three forms of this general type: Stem and Leaf Plots, Frequency Graphs, and
Histograms.
A "Stem and Leaf Plot" provides a graphic way to assemble or collect data
in the form of numbers. Many times a completed stem and leaf plot will give
at least some visual information about the data. In truth, a stem an leaf plot is
tabulation of data more than a graphic representation of the data. However, it
has a graphic element to it, and it can be seen as a primitive form of a histogram.
In its simplest form a stem and leaf plot is used on data consisting of two
digit numbers. In this case the "stem" of the graph is the …rst digit of the
number, and the "leaf" is the second. The stem numbers are most often stacked
vertically, and the leaves are placed to one side horizontally.

Consider our list of test scores.

April 55
Barry 63
Cindy 88
David 97
Eileen 58
Frank 90
Gena 88
Harry 71
Ivy 65
Jacob 77
Keri 75
Larry 88
Mary 95
Norm 86

If we are interested in the test more than the student’s scores, we rearrange the
table by dropping the names so that each datum is just a test score. All we
care about is the scores, so we rearrange the data by forgetting about the names
in these rows.

55; 63; 88; 97; 58; 90; 88; 71; 65; 77; 75; 88; 95; 86:

We end up with the most straightforward of all data sets: a simple list of
numbers.
Stem and Leaf Plots only work for data sets made up of numbers. Using

14
the …rst digit as the stem and the second digit as the leaf, we get a plot like

Stem Leaf
5 58
6 35
7 157
8 6888
9 057

We can easily see where Barry’s score of 63 appears in this plot, its stem is 6
and leaf is 3. Unfortunately, we no longer know it belongs to Barry. As we can
see, this works quite well on data sets made up of two digit numbers. But it is
easy to see how to adapt it to other types of numbers.
Consider the basic idea behind this stem and leaf plot. We start with a
data set that is either given or being collected. The datum in the set are single
objects, in the example, two digit test scores between 50 and 99. When we
chose to use the …rst digit of each number as a stem, we have actually broken
the data into classes: 50-59, 60-69, 70-79, 80-89, and 90-99. Once these classes
have been assigned as the stems, the rest of the datum became the leaf. Since
in the example the leaves are digits (numbers) we used the digits themselves to
represent the leaves, and arranged them in increasing order. When numbers are
involved, proper etiquette calls for the leaves to be listed in order.
This general description of a stem and leaf plot leads us to a "Class and
Frequency Plot." Data is divided into classes, and the classes play the role
of the "stems." Since the remainder of a datum may not be a digit, or even
a number, we might use any symbol to denote each leaf. Each piece of data
appears as a symbol in the correct stem. In our example of test data, we might
end up with something like

Stem Leaf
5
6
7
8
9

We de…nitely lose information this way, and this is probably not an example
where this is the best idea. However, there is information left that might be
useful. The number of leaves in each stem shows the frequency that data
occurs in this class. Thus the "Stem and Leaf Plot" has become a "Class and
Frequency Plot."
Consider a better example, namely the Eagles football team. Suppose that
we have collected data in the form (First Name, Last Name, Height, Weight.)
for all 80 players. We notice that the players are all between 68 and 80 inches
tall. We will assign every measurement in this range as a class. The classes refer
to only on part of each datum, and the rest of the datum is a bit complicated.

15
Still we will just denote the rest of each datum with a blue . The result is a
true class and frequency graph:

We easily see that 3 players are 77 inches tall.

If we tilt it on its side, it looks like

16
If we merge the diamonds into columns, we get

Notice all of these are class and frequency graphs drawn as columns or rows.
Thus class and frequency graphs can appear as bar charts and column charts.
In this example, the classes are the heights and are listed across one axis. The
columns measure a count of the number of times that a datum falls in a partic-
ular class; that is to say, it gives the frequency of that class in the whole.

Once we have a class and frequency chart, we are very close to our last type
of chart, a histogram. In a proper histogram, we also select classes, and separate
the data into those classes. Again we highlight the frequency with which each
class occurs in the data. In a histogram, we use a shaded area above a class
label - rather than a height - to indicate the frequency. Usually, but not strictly,
the scale marking the frequency is either a percentage or a proportion out of
1. To emphasize the area aspect of the graph, regions of the graph are run
together unless a class has a zero frequency of occurrence.

17
 Here we used the actually numbers to scale the frequencies. When the units
of the frequencies are uniform like this, there is no real distinction between the
height and area. Thus relabeling to a proportion of 1 does not change the shape
of the graph;

Prepared by: Daniel Madden and Alyssa Keri: May 2009

18