Chapter 5

Representation of Statistical Data

Dr Khalil Amine

[Link]@[Link]

2024 / 2025

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 1 / 54

 Representation using Tables

1. Representation using Tables

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 2 / 54

 Representation using Tables

1. Representation using Tables

A statistical table establishes a correspondence

between a set of individuals or modalities and one or
more corresponding data series composed of values,
frequencies, or counts.
X1 ... Xj ... Xp
ω1 x11 . . . . . . . x1p
.. ..
. .
ωi . . . . xij . . . .
.. ..
. .
ωn xn1 . . . . . . . xnp

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 3 / 54

 Representation using Tables Discrete Statistical Series

1.1. Discrete Statistical Series

Definition
Considering a statistical series (xi , ni )i=1,...,p ,
Count of the value xi is the number ni of times the value
xi appears.
∑
i
Cumulative count at xi is the sum ci = nj .
j=0

ni
Frequency of the value xi is the ratio fi = of the
N
count ni to the total count N of the population.
∑
i
Cumulative frequency at xi is the sum fc = fj .
j=0

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 4 / 54

 Representation using Tables Discrete Statistical Series

1.1. Discrete Statistical Series

Example
The iron contamination in the motor oil of 60 vehicles
was measured by burning a sample for each vehicle
and observing the emitted light. The obtained values
are as follows:

97 97 102 106 95 102 99 103 107 97 101 95 110 95 98

102 95 99 102 109 107 105 108 102 100 98 99 106 109 99
101 99 108 106 109 95 96 102 102 97 97 109 97 107 99
103 99 106 99 95 110 97 109 105 97 97 107 100 96 102

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 5 / 54

 Representation using Tables Discrete Statistical Series

1.1. Discrete Statistical Series

Example
These data can be presented in the following
statistical table:

Value 95 96 97 98 ... 108 109 110

Count ni 6 2 9 2 ... 2 5 2
Cumulative Count ci 6 8 17 19 ... 53 58 60
Frequency fi 0.10 0.03 0.15 0.03 ... 0.03 0.08 0.03
Cumulative Frequency fc 0.10 0.13 0.28 0.32 ... 0.88 0.97 1.00

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 6 / 54

 Representation using Tables Continuous Statistical Series

1.2. Continuous Statistical Series

Definition
Considering a statistical series (]ai , ai+1 ], ni )i∈A⊂Z ,
1. Count of ]ai , ai+1 ] is the number ni of times a value falls
within the interval ]ai , ai+1 ].
2. Cumulative count at ai is the total number ci of values
within the interval ]ainf A , ai ].
ni
3. Frequency of ]ai , ai+1 ] is the ratio fi = , where ni is the
N
count of the interval and N is the total count of the
population.
∑
i
4. Cumulative frequency at ai is the sum fc = fj .
j=0

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 7 / 54

 Representation using Tables Continuous Statistical Series

1.2. Continuous Statistical Series

Remark
The classes in a discrete or continuous series represent
modalities that can be of different types (e.g.,
([ai , ai+1 [, ni )i∈A⊂Z ) as long as they do not overlap.
Moreover, the classes in a continuous series are not
necessarily adjacent.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 8 / 54

 Representation using Tables Continuous Statistical Series

1.2. Continuous Statistical Series

Remark
The classes in a continuous series are not necessarily
of the same width. This depends on the studied case,
where class grouping must hold a certain significance.

For example, the acidity level (pH) can be measured

by grouping into three classes: ]0, 7[, {7}, and ]7, 14[,
forming a nominal grouping: Acidic, Neutral, and
Basic.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 9 / 54

 Representation using Tables Continuous Statistical Series

1.2. Continuous Statistical Series

Example
A blood test was conducted for 20 individuals. The
results for the white blood cell count (mm3 ) are as
follows:
3600 5100 6100 4600 5100 4800 3200
4700 5600 6900 5300 6100 3900 4200
6300 5400 6200 3200 3700 3400

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 10 / 54

 Representation using Tables Continuous Statistical Series

1.2. Continuous Statistical Series

Example
These data can be presented in the following
statistical table:

Class [3000, 4000 [ [4000, 5000[ [5000,6000[ [6000,7000[

Count ni 6 4 5 5
Cumulative Count ci 6 10 15 20
Frequency fi 30% 20% 25% 25%
Cumulative Frequency fc 30% 50% 75% 100%

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 11 / 54

 Graphical Representation

2. Graphical Representation

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 12 / 54

 Graphical Representation

2. Graphical Representation

The aim of graphical representation is to visually

summarise the information conveyed by the
observations collected for a variable.
The choice of representation depends on the nature of
the variables and ergonomic considerations.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 13 / 54

 Graphical Representation Case of a Qualitative Statistical Series

2.1. Case of a Qualitative Statistical Series

A qualitative statistical series can be represented

using two types of charts: bar charts for counts and
pie charts (Fr. camembert) for frequencies.

Figure: Bar chart and pie chart for qualitative data

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 14 / 54

 Graphical Representation Case of a Continuous Statistical Series

2.2. Case of a Continuous Statistical Series

2.2.1. Histogram:
A histogram consists of rectangles where the width
along the x-axis represents the class width, and the
height along the y-axis is such that the area of each
rectangle is proportional to the class count.

Figure: Histogram for continuous data

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 15 / 54

 Graphical Representation Case of a Continuous Statistical Series

2.2. Case of a Continuous Statistical Series

2.2.2. Frequency or Count Polygon:
The frequency or count polygon of a continuous
statistical distribution is obtained by joining the
midpoints of the upper horizontal sides in the
histogram of the distribution.

Figure: Frequency polygon for continuous data

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 16 / 54
 Graphical Representation Case of a Continuous Statistical Series

2.2. Case of a Continuous Statistical Series

2.2.3. Cumulative Frequency Polygon:
The cumulative frequency polygon of a grouped
statistical distribution graphically represents the
function defined over each interval ]ai , ai+1 ], 1 ≤ i ≤ p
by
∑
i−1
x − ai
f (x) = fj + fi .
j=1
ai+1 − a i

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 17 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.1. Stick Diagram:
The stick diagram for the counts (or frequencies) of a discrete
statistical distribution consists of a sequence of vertical segments
with x-coordinates xi , where the height is proportional to the
count (or frequency) of xi .

Diagram of iron contamination

Khalil Amine (EHTP, Casablanca)
measurements in motor oil.
[Link]@[Link] 18 / 54
 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.2. Frequency or Count Polygon:
The frequency (or count) polygon of a discrete
statistical distribution is derived from the stick
diagram by joining the tops of the segments with
straight lines.

Diagram of iron contamination

Khalil Amine (EHTP, Casablanca)
measurements in motor oil.
[Link]@[Link] 19 / 54
 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Definition
A quantile is any value that divides a statistical
series into adjacent intervals containing an equal
number of data points.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 20 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
The set of quantiles that are multiples of the fraction
q are called q-quantiles and consist of q − 1
1

quantiles, where q is a positive integer. The median is

an example of a quantile forming a set of 2-quantiles.
Quantiles allow for visualising the distribution of a
statistical series (sometimes called a dataset) on an
equally divided scale.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 21 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Definition
A quartile is any quantile that is a multiple of 14 .
The three quartiles, Q1 , Q2 , and Q3 , divide the
statistical series into four intervals containing an equal
number of data points (or individuals).

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 22 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
The second quartile, Q2 , corresponds to the dataset’s
median.

The first quartile, Q1 , corresponds to the median of

the values between the dataset’s minimum value and
Q2 .

The third quartile, Q3 , corresponds to the median of

the values between Q2 and the dataset’s maximum
value.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 23 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Remark
The following types of quantiles are commonly used:
Terciles: Two quantiles that are multiples of 13 .
Quintiles: Four quantiles that are multiples of 15 .
Percentiles: Ninety-nine quantiles that are
1
multiples of 100 .

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 24 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Example Consider the following dataset:
22 16 15 4 15 28 8 5 10
24 4 4 25 5 21 17 24 19
The three quartiles of this series are: Q1 = 5,
Q2 = 15.5, and Q3 = 22.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 25 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Definition
A boxplot (box-and-whisker plot) is a graphical
representation of the three quartiles, the minimum
value, and the maximum value of a statistical series.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 26 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Definition
The interquartile range (IQR) is the difference
between the third and first quartile: IQR = Q3 − Q1 .

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 27 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Definition
An outlier or an aberrant data is any value in a
statistical series that is significantly different from the
other observations.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 28 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Example
0 and 20 are two outliers in the following dataset:

8 8 0 11 9 9 8 8 8 10 8 11 9 9 14
9 11 12 9 13 9 8 13 14 14 9 10 8 8 14
20 8 8 12 11 11 12 12 11 8 11 12 12 9 9

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 29 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
The interquartile range provides a measure of data
dispersion that excludes the influence of outliers.
Indeed, the IQR represents the range of the central
50% of the dataset.

Remark
Outliers often indicate measurement errors and may
be excluded from statistical analyses in many
practices.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 30 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Exercise
Consider the following dataset:
18 22 21 22 25 14 18 12 23 15 14
23 23 12 24 22 24 20 24 16 17 19

1. Determine for this series: the median, the range,

the upper and lower quartiles, and the
interquartile range.
2. Draw a boxplot for this series.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 31 / 54

 Graphical Representation Case of a Discrete Statistical Series

2.3. Case of a Discrete Statistical Series

2.3.3. Boxplot (Box-and-Whisker Plot):
Exercise (Special Case)
Consider the following dataset: 3 7 14
1. Determine for this series: the three quartiles, the
range, and the interquartile range.
2. Draw a boxplot for this series.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 32 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.1. Scatter Plot:
Given two statistical variables X1 and X2 observed in
a population of size N , a scatter plot of individuals is
a set of points in the space R2 representing the profiles
(numeric pairs) (X1i , X2i ) for i = 1, . . . , N , where Xji for
j ∈ {1, 2} is the value of variable Xj for individual i.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 33 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.1. Scatter Plot:
The scatter plot allows us to visualise the distribution
of profiles (e.g., response profiles, sensory profiles,
measurement profiles, etc.) in the(space and) their
dispersion around a mean profile X 1 , X 2 .

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 34 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.1. Scatter Plot:

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 35 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Fitting the scatter plot involves determining a
relationship between the two variables X1 and X2 .
The aim is to find a function whose graph closely
aligns with the points of the scatter plot.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 36 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Linear Fit

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 37 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Logarithmic Fit

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 38 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Exponential Fit

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 39 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Polynomial Fit

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 40 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Remark
Observe the distinction between numerical
interpolation and the types of fits illustrated in the
above figures.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 41 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
In general, given two variables X and Y , fitting is
part of a broader set of techniques called regression,
which seeks to find a function f such that Y ≈ f (X ),
along with a criterion to quantify the quality of the fit
of f to the data.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 42 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
The aim of simple regression (resp. multiple
regression) is to express a variable Y , known as the
dependent variable, in terms of an independent
variable X (resp. multiple independent variables
X1 , . . . , Xq ).

Regression applies when the relationship between X

and Y is more than mere correlation and represents a
causal connection.

For multiple regression, the variables Xi for

i = 1, . . . , q must be independent.
Khalil Amine (EHTP, Casablanca) [Link]@[Link] 43 / 54
 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Suppose Y can be ex-
pressed in terms of X by
the equation y = ax + b.
For each point Mi (xi , yi ) in
the scatter plot, a point Pi
on the line D can be as-
sociated with the same x-
coordinate as Mi .

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 44 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
The deviations between
the points Mi in the scat-
ter plot and the line D are
given by:
∑ ∑
2
(Mi Pi ) = (yi −axi −b)2

This method is called the

least squares method.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 45 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Theorem
∑ ∑
The sum (Mi Pi )2 = (yi − axi − b)2 is minimised
for:
cov(X , Y ) cov(X , Y )
a= , b=Y − X
var(X ) var(X )

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 46 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Theorem
The regression line of a statistical variable Y on a
statistical variable X is given by the formula:
cov(X , Y ) cov(X , Y )
y= x +Y − X
var(X ) var(X )

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 47 / 54

 Graphical Representation Case of Two Statistical Series

2.4. Case of Two Statistical Series

2.4.2. Fitting a Scatter Plot:
Exercise:
Given the following statistical series:
Year 1 2 3 4
Sales in KDH 4 9 11 12

1. Represent the series graphically.

2. Determine the regression line using the least
squares method.
3. Plot the regression line on the same graph.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 48 / 54

 Graphical Representation Case of Multiple Statistical Series

2.5. Case of Multiple Statistical Series

2.5.1. Box Plot:
A box plot is a graphical representation used to
compare quantitative statistical series with different
sample sizes, such as the case of observing the grades
of four classes.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 49 / 54

 Graphical Representation Case of Multiple Statistical Series

2.5. Case of Multiple Statistical Series

2.5.2. Correlation Circle:
A correlation circle is a graphical representation of the
projection of correlations between multiple statistical
variables on a circle of radius 1 in the space R2 , where
the distance between two variables represents the
degree of correlation between them.

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 50 / 54

 Graphical Representation Case of Multiple Statistical Series

2.5. Case of Multiple Statistical Series

2.5.2. Correlation Circle:

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 51 / 54

 Exercise

3. Exercise

Khalil Amine (EHTP, Casablanca) [Link]@[Link] 52 / 54

 Exercise

3. Exercise

A primary school teacher has collected the grades

obtained by their students in the last test and
presents the results in the form of the box plot below:

1. What information can be extracted from this

graphical representation?
Khalil Amine (EHTP, Casablanca) [Link]@[Link] 53 / 54
 Exercise

3. Exercise

2. What conclusions can be drawn from the

following situations?
Ayoub scored 2.5
Youssef scored 8.5
Meryem scored 4.5
Jad scored 6.5
Khalil Amine (EHTP, Casablanca) [Link]@[Link] 54 / 54