Graphical Descriptive Tools

1 Introduction
One of the areas that descriptive statistics deals with is presen-
tation of data in various forms. We have graphical descriptive
tools and non-graphical descriptive tools. Graphical descriptive
statistics are the various graphs, tables and diagrams used to dis-
play main features of a given dataset. Non-graphical statistics are
numbers that summarise an aspect of the data. In short, one of
the uses of descriptive statistics is to obtain a summary of a given
data set.

We use descriptive tools to

• simplify and condense data,
• present facts,
• do comparisons, and
• relate the variables.
Data analysis usually begins with data exploration which makes
use of descriptive tools. The goal is to develop a high-level under-
standing of the data at hand.

Tools appropriate for one task might not be appropriate for an-
other. In some cases some tools perform better than other tools.
As a scientist, the choice depends on the kind of data and the task
to perform.

1
2 Intended Learning Outcomes
At the end of this lesson, a student is expected to be able to
• describe what graphical descriptive tools are,
• differentiate between graphical descriptive tools and non-
graphical descriptive tools,
• present data using an appropriate graphical descriptive tool,
and
• read information presented by a graphical descriptive tool.

3 Tabular Summaries
There are various ways of organising and presenting data. Statis-
tical tables are the simplest and most revealing devices for sum-
marising and presenting data. Data are organised in rows and
columns. In some cases, it gives us the how often each value ap-
pears in a given data set- the frequency. Such a table is called a fre-
quency table or frequency distribution- or simply the distribution.
We use tables when handling both qualitative and quantitative
data.

3.1 One-Way Table

With this kind of table, we summarise a data set based on only
one characteristic.

2
Example 1

There is an outbreak of a disease in a village. A sample of 20

individuals is tested for the disease. A D is used to mark every
test that gives a positive result, and a U is used to mark every
test that gives a negative result.

D, D, U , D, U , D, D, U , U , U
D, D, D, D, U , D, U , D, D, D

Required to: summarise the dataset using a table.

Presenting data as it is will make it difficult for the audience to

grasp quickly information hidden in the data set. A quick way to
present this information to the audience can be through a table.
Since we are classifying the data based only one characteristic,
test result, we will use a one-way table.

Test result Number of patients

Detected 13
Undetected 7
Total 20
Table 1: Distribution of test results.

From the table we can quickly tell that there may be more
infected people (at least 50%) than the uninfected ones in the
village.

3
Example 2

Consider the number of patients attending physiotherapy ses-

sions at a rehabilitation center- the data was recorded for a
period of 30 days.

2, 2, 0, 3, 3, 1, 3, 4, 6, 5
0, 1, 1, 5, 5, 1, 1, 4, 2, 2
3, 3, 2, 4, 4, 3, 3, 4, 4, 3

Required to: summarise the dataset using a table.

We can organise this data in a one-way table as below (we are

classifying the data based on only one characteristic, number of
patients).

Number of patients Number of days

0 2
1 5
2 5
3 8
4 6
5 3
6 1
Total 30

4
 Table 2: Distribution of number of patients attending physio-
therapy sessions.

From the table it is easier to tell that on most days 3 patients

attended the sessions. The minimum number of patients at-
tending is 0 and the maximum is 6.

3.2 Two-Way Table

With this kind of table, data is classified based on two character-
istics.

Example 3

Let us look at Example 1 one more time- we are looking at

an outbreak of a disease in a village. Suppose in the study
gender was also considered. The colour blue was used for a
male individual and black for a female individual.

D, D, U , D, U , D, D, U , U , U
D, D, D, D, U , D, U , D, D, D
Required to: summarise the dataset using a table.

Since we are classifying our data based on test result (positive

or negative) and gender (male or female), we will use a two-way
table- we are using two attributes to classify the data.

5
 Test Result
Gender Detected Undetected Total
Male 5 3 8
Female 8 4 12
Total 13 7 20
Table 3: Distribution of test results.

From the table it is easier to tell that five out of the eight males
tested positive, and eight out of the 12 females tested positive.
The proportion of males who were positive might be lower than
that of females.

3.3 Continuous Data

With continuous data, in many cases the number of unique values
is high. Constructing a frequency table using the raw values results
in a ’summary’ that has too much detail- our aim is to summarise
the data set. Such data can be grouped. The created groups are
supposed to be non-overlapping. For each group a frequency is
obtained.

1. Divide the data set into k intervals starting with lower bound-
ary y1 and ending with upper boundary yk+1. y1 can be the
smallest value in the data set or slightly smaller than the
smallest value; and yk+1 needs to be slightly larger than the
largest value in the data set. k can be the smallest value
such that 2k > n, where n is the size of the data set.
yk+1 −y1
2. (The class width is given by w = k . When the result
6
 is containing a fraction part, round up the value. When we
add class width to lower boundary of a class, we get the
upper boundary of that class.)
3. The first interval is from y1 inclusive to y2 exclusive; the
second interval is from y2 inclusive to y3 exclusive; the third
interval is from y3 inclusive to y4 exclusive;...; the kth inter-
val is from yk inclusive to yk+1 exclusive.
4. Group intervals are not necessarily supposed to be equal. w
could be different for different groups.
5. Obtain the frequency of each group interval.
6. Display the group intervals and their frequencies/frequency
densities in a table.
Group interval Frequency
y1≤x<y2 f1
y2≤x<y3 f2
y3≤x<y4 f3
.. ..
yi≤x<yi+1 fi
.. ..
yk ≤x<yk+1 fk

Example 4

Below are heights in centimetres of some children of a certain

7
age- the data was being used in a pulmonary function related
study.

82, 83.5, 85.7, 87, 87, 88.2, 89.1,

89.3, 90.1, 91.1, 92, 93, 93, 93.2,
93.5, 93.5, 93.9, 94, 94.3, 95.7, 96,
96, 97, 99, 100.6, 100.9, 101, 103,

Required to:

• determine the number of group intervals for the dataset.

• determine the class width.

• determine the group intervals (classes).

• provide a summary of the dataset using a table.

There are 28 observations (n = 28). To determine the number

of group intervals to have, we will use the relationship 2k > n
where k is the required number of group intervals. Thus the
smallest value of k that gives us 2k > 28 is 5- we will have 5
group intervals.

We choose the minimum value to be 80 and the maximum value

to be 105. Then class width will be w = xmax−x
k
min
= 105−80
5 = 5.

Our classes will be

80 − 85, 85 − 90, 90 − 95, 95 − 100, 100 − 105.

Then the table summarising our data set is below.

8
 Height (cm) Frequency
80 − 85 2
85 − 90 6
90 − 95 11
95 − 100 5
100 − 105 4
Total 28
Table 4: Distribution of heights of children in a pulmonary
function related study.

From the table, we may say that there were more children with
heights in the interval 90 − 95 (in cm) in the study. The interval
80 − 85 (in cm) had the least number of children.

In some cases relative frequency is used- it is given by sum frequency

of frequencies .
(When it comes to the frequentists, long run relative frequency is
probability. Relative frequency can also be used when handling
the other types of variables- not continuous data only.)

Example 5

Let us consider Example 4. The table will look like the one
below when we use relative frequencies.

9
 Height (cm) Relative Frequency
1
80 − 85 14
3
85 − 90 14
11
90 − 95 28
5
95 − 100 28
1
100 − 105 7
Total 1
Table 5: Distribution of heights of children in a pulmonary
function related study.

For example, if we consider the group 85 − 90 which has the

3
relative frequency 14 , we can say that for every 14 children there
are 3 children whose height is between 85cm and 90cm (85cm
is inclusive while 90cm is not).

Group intervals are not necessarily supposed to have the same

width- it could be different for different groups. In such a case
we use frequency density, and not frequency or relative frequency.
frequency for that class
Frequency density for a given class is class width for that class . Thus
multiplying the class width and the frequency density will give us
the frequency.

Example 6

Suppose in Example 4 we use the intervals 80 − 87, 87 − 90,

90 − 95, and 95 − 105. The widths are different. We will use
frequency densities as below.

10
 Height (cm) Frequency Frequency Density
80 − 87 3 0.429
87 − 90 5 1.667
90 − 95 11 2.2
95 − 105 9 0.9
Total 28
Table 6: Distribution of heights of children in a pulmonary
function related study.

3.4 Cumulative Frequency Distributions

The less than cumulative frequency distribution gives the number
of observations less than a specified value, and the more than cu-
mulative frequency distribution gives the number of observations
more than or equal to a specified value.

Example 7

Consider the data in Example 4.

Required to:
• come up with the ’less than’ cumulative frequency distri-
bution for the dataset.
• come up with the ’more than’ cumulative frequency dis-
tribution for the dataset.

11
In the ’less than’ cumulative frequency distribution, the fre-
quencies for the intervals 80 − 85, 85 − 90, 90 − 95, 95 − 100,
and 100−105 are 2, 2+6 = 8, 2+6+11 = 19, 2+6+11+5 = 24
and 2 + 6 + 11 + 5 + 4 = 28 respectively. We have the table
below.
Height (cm) Cumulative Frequency
80 − 85 2
85 − 90 8
90 − 95 19
95 − 100 24
100 − 105 28
Table 7: ’Less than’ cumulative frequency distribution of
heights of children in a pulmonary function related study.

From the table we can easily tell that, for example, there were
19 children with a height less than 95cm in the study.

Using Example 4 let us look at the ’more than’ cumulative

frequency distribution for the data. the frequencies for the in-
tervals 80 − 85, 85 − 90, 90 − 95, 95 − 100, and 100 − 105 are
4 + 5 + 11 + 6 + 2 = 28, 4 + 5 + 11 + 6 = 26, 4 + 5 + 11 = 20,
4 + 5 = 9 and 4 respectively.

Height (cm) Cumulative Frequency

80 − 85 28
85 − 90 26
90 − 95 20
95 − 100 9
100 − 105 4

12
 Table 8: ’More than’ cumulative frequency distribution of
heights of children in a pulmonary function related study.

From the table we can tell that, for example, there were 20
children with height more than or equal to 90cm.

It is crucial that you know how to interpret correctly these tables,

and the graphical tools that we will look at later.

4 Graphical Summaries
It is said that ’a good picture is worth 1000 words!’ Graphs are
used in getting a quick overall impression. Sometimes they are
used alongside tables. Graphs are used in summarising a given
data set- the aim is to provide summary of an attribute of interest
using a picture.

4.1 Pie Chart

It is a commonly used form of diagram. It is useful in showing
how a total amount is divided into constituent parts. It is a way
of summarizing a set of categorical data or displaying the different
values of a given variable. It is used in comparing the categories.
Steps provided below may be followed in constructing a pie chart.

1. Identify the categories (or unique values) in the data set.

Divide the data into the categories.
2. Construct a frequency table, showing the categories (or unique
values) and their frequencies.

13
 3. Convert the frequencies into angles.
4. Construct a circle and split it into sectors using the angles
obtained in the preceding step.
5. Label the sectors.

When the differences in the quantities of the groups of interest are

small, comparison using a pie chart without supporting numbers
is not an easy task.

Example 8

The content for a package of crisps is 6.1g, 34.2g, 48.1g and

11.6g (per 100g) for protein, fat, carbohydrates, and dietary
fibre respectively.

Required to: present the given data using a pie chart.

We can present this using a table, but we will go for a pie chart
as required. We will convert the contents into angles. We have
6.1 ◦ ◦ 34.2 ◦ ◦ 48.1 ◦ ◦
100 ×360 = 21.96 , 100 ×360 = 123.12 , 100 ×360 = 173.16 ,
◦ ◦
and 11.6
100 ×360 = 41.76 for protein, fat, carbohydrates, and
dietary fibre respectively.

Then we can have our pie chart.

14
 Figure 1: Content of a package of crisps.

From the pie chart one can quickly tell that the package contains
more carbohydrates compared to the other nutrients. Protein
is the least.

4.2 Bar Graph

This is another commonly used tool. Just like the pie chart, the
bar chart is a way of summarizing a set of categorical data or
displaying the different values of a given variable. It is used in
comparing the categories. The steps below may be followed in
constructing a bar graph.
1. Identify the categories (or unique values) in the data set.
Divide the data into the categories.
2. Construct a frequency table, showing the categories (or unique
values) and their frequencies.
3. For each group construct a rectangle with constant width
and height proportional to frequency.
4. Rectangles are contiguous- there are spaces between rectan-
gles.
15
We have simple, component, and multiple bar graphs.

Example 9

We will consider the data provided in Example 8.

Required to: present the given data using a bar graph.

Figure 2: Content of a package of crisps.

From the bar chart we can quickly have the information of the
content of the package- for example, the least is protein.

Example 10

In a study, the focus was on the performance of a test with

16
regard to a particular disease. Out of 20 patients suffering
from the disease, for 13 patients the test came positive while
for 7 it was negative. Out of the 13 with a positive result
5 were male, and out of the 7 with a negative result 3 were male.

Required to: present the given data using a bar graph.

We are classifying our data based on test result and gender- thus
a two-way table is possible. But we will produce a multiple bar
graph.

Figure 3: Distribution of test results.

We can quickly tell that for both males and females the test
might be at least 50% correct.

17
Example 11

We will look at number of cases of a disease observed at three

hospitals- Hospital A, Hospital B, Hospital C- in August,
September, October, November, and December last year.
Hospital A observed the following numbers of cases: 2, 9, 3,
11, and 9 (respectively); Hospital B observed the following
numbers of cases: 4, 8, 7, 3, and 12 (respectively); and Hospital
C observed the following numbers of cases: 5, 2, 8, 10, and 11
(respectively). The totals for these months are 11, 19, 18, 24,
and 32 respectively.

Required to: present the given data using a bar graph.

Here we are classifying our data based on hospital and month of

the year. We can use a multiple bar graph- but suppose our in-
terest is in showing totals for each month and components/parts
forming the total. Then we will use a component (stacked) bar
graph- a table can do the work, but we will go for a graph.

18
 Figure 4: Bar graph for a disease (cases) at Hospital A, Hospital
B and Hospital C in August, September, October, November
and December.

The graph tells us that the highest number of cases was observed
in December.

4.3 Scatter Plot

A scatter plot is used in visualising the relationship between two
variables that are recorded from the same subject/item. The ob-
servations of the two variables are paired. One variable is placed
on the horizontal axis, while the other variable is on the vertical
axis. The pairs are plotted as points. The scatter plot gives the
form of the relationship- for example, linear or non-linear; the di-
rection of the relationship- positive or negative; and the strength
of the relationship- weak or strong.

19
Example 12

In a study the researcher was looking into the relationship be-

tween fat (grams) and total calories. Different sandwiches were
considered.

20
 Total Fat (g) Total Calories
9 260
13 320
21 420
30 530
31 560
31 550
34 590
25 500
28 560
20 440
5 300
Table 9: Fat and total calories for different sandwiches.

Required to:

• present the given data using a scatter plot.

• describe the relationship between total fat and total calo-

ries.

Looking at the table above it will be hard to quickly provide

a good explanation on the relationship that is existing between
fat (grams) and total calories. A scatter plot may do the job
better.

21
 Figure 5: Fat grams and total calories for different sandwiches.

We may say that there is possibly a linear relationship between

total fat and total calories. It seems to be a strong positive
relationship. As total fat increases, total calories increases too
(and vice versa).

4.3.1 Line Graph

A line graph, or line chart, is used in observing the trend over
time- it shows how a variable changes over time. It works well with
continuous data- in some cases interpolation can be performed.

Example 13

A personal doctor has been following the health of a patient for

a period of eight years. One of the quantities s/he has been

22
recording is average body mass (kilograms) of the patient. The
table below is the data.
Year Body mass (kg)
2006 75
2007 79
2008 83
2009 86
2010 85
2011 83
2012 87
2013 89
Table 10: Average body mass of a patient.
Required to: present the data using a line graph.

A quick way to observe the changes in body mass with respect

to time is a line graph- a table may do but it will take time to
grasp the information being presented.

Figure 6: Average body mass of a patient.

23
Maybe if you were the doctor you might have an explanation
for a steady increase in body mass for years from 2006 to 2009,
and the drop from 2009 to 2011.

Example 14

In a first mathematics test that a class wrote, on average

girls scored 61%; in the second test the average score was
75%; in a third test it was 72%; and in a fourth test it was
88%. The average score of boys in similar tests were 79%,
80%, 85% and 84% respectively. The period between the
tests was the same. The first test was written at the start of
the term while the fourth test was written at the end of the term.

Required to: present the data using a line graph.

We can visualise the performance of the girls and boys using

graphs to have an idea of how the scores were changing. The
line graphs/charts below show the scores.

24
 Figure 7: Line graphs of average scores of boys and girls in four
different mathematics tests.

What seems to be the increase in performance of girls is easily

noticeable than that of boys- the improvement on the side of
girls seems to be much compared to that of boys. The perfor-
mance of boys in the first three tests seems to be better than
that of girls.

4.4 Histogram
A histogram is a convenient way of displaying a frequency table.
It presents a graphic picture of the distribution of measurements.
This picture consists of rectangular bars joining each other, one
for each group interval. A histogram is for continuous data- it
gives the shape of the distribution.

25
 (a) Symmetric. (b) Left skewed. (c) Right skewed.
Figure 8: Shapes of distributions.

If we draw a vertical line in the middle of our histogram, with a

symmetric distribution the number of observations (data points)
on either side of the vertical line is the same; for a left-skewed
distribution there are more observations to the right side of the
line than the left- the tail is long to the left; and for a right-skewed
distribution there are more observations to the left side of the line
than the right- the tail is long to the right.

Example 15

Consider heights of children in a pulmonary function related

study provided in the table below.

26
 Height Frequency
80 − 85 2
85 − 90 6
90 − 95 11
95 − 100 5
100 − 105 4
Total 28
Table 11: Distribution of heights of children in a pulmonary
function related study.
Required to:
• present the distribution using histogram.
• determine the shape of the distribution.
For the above data a table as above may do but the audience
will not quickly figure out the shape of the distribution- tally
marks may help. A histogram may be a better way to go.

Figure 9: Distribution of heights of children in a pulmonary

function related study.

27
 For this study, we might conclude that the heights were approx-
imately symmetric (slightly skewed).

4.5 Box-And-Whisker Plot

Box-and-whisker plot, just like the histogram, is used in looking
at the shape of the distribution of measurements. It uses the re-
lationship among the median Q2, lower quartile Q1, and upper
quartile Q3 to give the shape of the distribution.

Since we have not yet looked at measures of central tendency and

dispersion, we will try to keep this simple in order to avoid most
of such computations- and keeping in mind that most of this work
these days is done by the computer. We are also not going to look
at outliers.

How do we obtain Q1, Q2 and Q3? We first need to arrange the

data set in ascending order. Then obtain the median, Q2. Q2
splits the data set into two subsets each with about 50% of the
observations. Thus one subset contains values that are less than
or equal to Q2, and the other contains values that are greater than
or equal to Q2. Then we need to obtain the median of the subset
containing values that are less than or equal to Q2- this median
is Q1 (lower/first quartile). We also need to obtain the median of
the subset containing values that are greater than or equal to Q2-
this median is Q3 (upper/third quartile).

28
• If the shape is symmetric, then the lower quartile and the
upper quartile are supposed to be the same distance from
the median- we have Q2 − Q1 = Q3 − Q2.

LV Q1 Q2 Q3 HV

• If the lower quartile is further away from the median than the
upper quartile, then we have negative skewness (left skewed);
Q2 − Q1 > Q3 − Q2.

LV Q1 Q2 Q3 HV

• If the upper quartile is further away from the median than

the lower quartile, then we have positive skewness (right
skewed); Q2 − Q1 < Q3 − Q2.

LV Q1 Q2 Q3 HV

29
Example 16

Let us consider the example that we have already looked at- the
heights of children of a particular age in a pulmonary function
related study.

82, 83.5, 85.7, 87, 87, 88.2, 89.1,

89.3, 90.1, 91.1, 92, 93, 93, 93.2,
93.5, 93.5, 93.9, 94, 94.3, 95.7, 96,
96, 97, 99, 100.6, 100.9, 101, 103,

Required to:

• present the distribution of the data using box-and-whisker

plot.

• determine the shape of the distribution.

We have already explored the shape of the distribution using

the histogram. For the sake of learning, let us try the box-and-
whisker plot.

The lowest value is 82 and the largest is 103. The median is

Q2 = 93.2+93.5
2 = 93.35.

The median of 82, 83.5, 85.7, 87, 87, 88.2, 89.1, 89.3, 90.1,
91.1, 92, 93, 93, and 93.2 is Q1 = 89.1+89.3
2 = 89.2.

30
 The median of 93.5, 93.5, 93.9, 94, 94.3, 95.7, 96, 96, 97, 99,
100.6, 100.9, 101, 103 is Q3 = 96+96
2 = 96.

We will plot the values of lowest number, Q1, Q2, Q3 and largest
number, then come up with the diagram as below.

Figure 10: Distribution of heights of children in a pulmonary

function related study.

The distribution seem to be slightly skewed to the left- we may

say it is almost symmetric. (Also we have Q2 − Q1 = 93.35 −
89.2 = 4.24 and Q3 − Q2 = 96 − 93.35 = 2.65; thus Q2 − Q1 >
Q3 − Q2.)

When it comes to the histogram and the box-and-whisker plot,

histograms are preferred when there is wide or little variation
among the observed frequencies. When there is moderate vari-
ation among the observed frequencies the box-and-whisker plot is
the preferred tool. Size of the data set is also one of the things we
need to consider when using these tools.

4.6 Stem-and-Leaf Diagram

Stem-and-leaf diagram is used in exploring the shape of the distri-
bution. The diagram is formed by splitting the values into stems
31
and leaves. Stem-and-leaf diagram retains the individual values.
It is convenient when working with small samples.
• Separate each data point into a stem component and a leaf
component. The stem consists of the number formed by all
but the leftmost digit of the number, and the leaf consists
of the rightmost digit.
• The smallest stem is displayed in the upper part of the plot.
• The second stem is the first stem plus 1 and it is displayed
below the first stem.
• The preceding step is repeated until the largest stem is
reached.
• A vertical bar is placed to the right of stems.
• For each number, the appropriate stem is found and the
leaves are displayed to the right of the vertical bar.

32
Example 17

The data considered in this example are weights in pounds of

children at a day-care center.

23, 25, 44, 12, 43, 25, 19, 49, 36, 36, 28, 31,
31, 42, 38, 21, 25, 24, 28, 47, 23, 22, 50, 27,
28, 16, 23, 65, 57, 49, 12, 32, 12, 38, 42, 27,
51, 63, 42, 27, 30, 28, 32, 43, 79, 27, 22, 24,
68, 69, 45, 46, 30, 74, 43, 51, 49.

Required to:
• present the distribution of the data using stem-and-leaf
diagram.
• determine the shape of the distribution.
In this example, the tens are going to form the stem components
while the units (ones) will form the leaf components- for exam-
ple, for 23 the stem component is 2 and the leaf component is 3;
for 25 the stem component is 2 and the leaf component is 5 e.t.c.

The stem-and-leaf diagram is provided below.

33
 1 22269
2 1223334455577778888
3 0011226688
4 2223334567999
5 0017
6 3589
7 49
Figure 11: The distribution of weights (pounds) of children at
a day-care center.

We may say that the weights are right-skewed.

4.7 Cumulative Frequency Polygon

Under the subsection Cumulative Frequency Distributions we con-
sidered the less than and the more than cumulative frequency
distributions- we presented these distributions using tables. These
distributions can be presented using graphs called cumulative fre-
quency polygons. Polygons are points on the graph, with the
successive ones joined by straight lines.

Example 18

We will consider the cumulative frequency distributions given

in Example 7- we are considering heights of children in a pul-
monary function related study. For the less than cumulative

34
frequency distribution the intervals 80 − 85, 85 − 90, 90 − 95,
95 − 100 and 100 − 105 have the cumulative frequencies 2,
8, 19, 24, and 28 respectively. The polygon for the less than
cumulative frequency distribution is below.

Required to:

• present the ’less than’ cumulative frequency distribution

using a polygon.

• present the ’more than’ cumulative frequency distribution

using a polygon.

Figure 12: ’Less than’ cumulative distribution of heights of

children in a pulmonary function related study.

From the figure we can tell that, for example, 19 children have
their heights less than 95cm.

We also considered the more than cumulative frequency dis-

tribution in the same example- the intervals 80 − 85, 85 − 90,

35
90 − 95, 95 − 100 and 100 − 105 have the cumulative frequencies
28, 26, 20, 9, and 4 respectively. The polygon for the more than
cumulative frequency distribution is below.

Figure 13: ’More than’ cumulative frequency distribution of

heights of children in a pulmonary function related study.

Thus we have, for example, 26 children with a height of more

than or equal to 85cm.

Example 19

Depending on orders, the numbers of an assistive device of a

certain type produced by a company were 5, 7, 4, 4, 3, 5, 0, 3, 0,
2, 5 and 4 for the months of January, February, March, April,
May, June, July, August, September, October, November
and December respectively. The company is interested in the

36
cumulative figures of the device they had produced.

Required to: present the cumulative frequency polygon.

The cumulative numbers for the respective months are 5, 5+7 =

12, 5 + 7 + 4 = 16, 5 + 7 + 4 + 4 = 20, 5 + 7 + 4 + 4 + 3 = 23,
5 + 7 + 4 + 4 + 3 + 5 = 28, 5 + 7 + 4 + 4 + 3 + 5 + 0 = 28,
5+7+4+4+3+5+0+3 = 31, 5+7+4+4+3+5+0+3+0 = 31,
5+7+4+4+3+5+0+3+0+2 = 33, 5+7+4+4+3+5+0+3+
0+2+5 = 38, and 5+7+4+4+3+5+0+3+0+2+5+4 = 42.
Below we have the polygon.

Figure 14: Cumulative distribution for the production of the

device.

By the end of the month of June the company had produced 28

devices. Between the months of June and July, and the months
of August and September the line is horizontal- telling us that
there was no production of the device in the months of July and

37
 September. (Remember that a table can be used for the same
task we have done here.)

For the tools that we have considered, it is necessary to provide a

caption. In some cases, the caption is provided alongside the title.
For the table we may be required to provide row and/or column
headings- in some cases totals are required. Figures in three di-
mension are not encouraged- make your figure easily understood
by your audience. Where the differences between/among groups
are very small, a pie chart that does not give any indication of the
amounts for each group is not a suitable tool.

38