MAKERERE UNIVERSITY BUSINESS

BUSINESS STATISTICS
TIME SERIES ANALYSIS
Outline of content
Introduction
Definition, description and examples
 Importance of time series
Components of time series
Decomposition of time series
Measuring trends
Measuring seasonal indices

Time series notes 10/09/2024 1

 Introduction
• When data is arranged according to the time of occurrence, it forms a
time series.
• Thus time series is the arrangement of statistical data in
chronological order i.e. daily, weekly, monthly, annually, semi-
annually, quarterly etc.
• Time series can also be defined as data collected sequentially over
time i.e. daily, weekly, monthly, annually, semi-annually, quarterly
etc.
• Examples of time series data include; annual profits of a company,
weekly sales, monthly output, annual operational costs, commodity
prices, employee wages etc.
• The series vary from one time period to another
Time series notes 10/09/2024 2
 Importance of time series
• Time series play an important role in the following ways
 Planning
 Forecasting
 Predicting/ forecasting trends
 Evaluating performance
 Policy formulation
 Decision making
 Comparison
Time series notes 10/09/2024 3
 Components of time series

• When data is arranged on the basis of time of their

occurrence, often they show fluctuations and this can be
explained by the four components of time series i.e.
 Secular trend
 Seasonal variations
 Cyclical variations
 Irregular/ random fluctuations

Time series notes 10/09/2024 4

 Trend/Secular trend
• This refers to the tendency of the series to either increase/ raise
or decrease/fall over along period of time
• Trend is that irreversible movement in time series which
continues in general in the same direction over long time period
• When series are increasing, then it is an upward trend and if it
is falling then it is a downward trend
• Upward trend;- population, technological advancement etc.
• Downward trend;- illiteracy levels
• Helps in studying patterns and behaviors of a phenomenon or
compare 2 or more time series over time
Time series notes 10/09/2024 5
 Illustration Upward and downward trend
Upward trend Downward
trend

Populati
Illiteracy
on

Time
Time series notes Time 10/09/2024 6
 Seasonality
• Seasonal variations are
Seasonali
variations/ movements that
ty
occur regularly every year
• Occurs in a regular and periodic
manner over a span of less
than a year e.g. agricultural
production
• Mainly caused by climate and
customs (habits and fashions)
• Helps in planning future options

Time series notes 10/09/2024 7

 Cyclical variations
• These are wavelike movements in Business
the time series with a period of cycle
oscillation of more than one year
• Most businesses and economies
experience these kind of variations
for example a boom an a depression
• During a boom, production, sales,
employment and profits among
others are high
• The reverse with the depression
• Helps in policy formulation aimed at
stabilizing the level of economic
activities
Time series notes 10/09/2024 8
 Random/ irregularity
• Theseare series which may irregularity
take a very long time to
occur
• May occur just once, or
several times but with out a
pattern or regularity
• Include wars, floods, earth
quakes , strikes etc.
• May be planned (episodic)
or un planned ( accidental)
Time series notes 10/09/2024 9
 Decomposition of time series

• There are 2 models that are used in the decomposition of time

series into its components i.e.
• Additive and multiplicative model
• Additive model: a time series is a sum of its 4 components.
Symbolically,
• where
• The model assumes that the all the components are
independent of one another i.e. T has no effect on S and C nor
does S have an effect on C and vice versa
• Different components are absolute quantities expressed in
10
original units i.e can take both negative and positive values
Time series notes 10/09/2024
 Decomposition of time series
• A time series is a product of its 4 components. Symbolically,
•
• In this model, its only trend represented in terms of original
values while S and C are expressed in relatives/ percentages
• Assumes that the components are due to different causes so
not necessarily the independent
• Most series conform to the multiplicative model
Time series notes 10/09/2024 11
 Measurement of trend
• There are several method that can be used in measuring trend for
example the freehand method, the selected points method, the
semi-average method, the method of moving averages and least
squares among others
• In this course, we shall mainly focus on the least squares
method
• By using the least square method, the trend line/equation is fit
• In some literature a trend equation is used but to avoid confusion
with the regression model, we shall stick on the equation

• Where .
Time series notes 10/09/2024 12
 Measurement of trend

Example One:
The data below relates to the enrollments to the BCOM evening group
for the past 9 years. Use it to predict the enrollment rate for the year
2025
Year 2015 2016 2017 201 2019 2020 2021 2022 2023
8
No of 90 100 80 85 105 90 100 86 98
students

Time series notes 10/09/2024 13

 Solution
odd number of years (computing t)

Year y t ty
2015 90
2016 100
2017 80
2018 85
2019 105
2020 90
2021 100
2022 86
2023
Time series notes 98 10/09/2024 14

Total 834
 Solution

From
• Where .

Time series notes 10/09/2024 15

 Measurement of trend

Example Two:
The data below relates to the annual budget of Antonio
Year
investments 2016 2017
for the past201 2019
8 years. 2020 2021 2022 2023
8
Budget 100 80 85 105 90 100 86 98
(million)

Required;
a) Fit the trend equation
b)Determine the annual budget for Antonio investments in the
yearTime
2026.
series notes 10/09/2024 16
 Solution
Even number of years (computing t)
Year y t ty
2016 100
2017 80
2018 85
2019 105
2020 90
2021 100
2022 86
2023 98
Total 744
Time series notes 10/09/2024 17
 Solution

• 96

• The annual budget will be 96 millions in the year 2023

Time series notes 10/09/2024 18

 Trial Question
The management of sole Candidature Enterprise Ltd is currently faced with a
challenge of predicting what the sales will be in 2022. Management has approached
you for assistance. Inform management what the sales will be using the information
below
Year 2011 2012 2013 2014 2015 2016 2017 2018 2019 202
0
Adv. Expenditure 11 10 12 15 15 23 24 38 42 48
(millions)
Sales Income 3 4 5 5 6 7 8 9 10 11
(millions)
b) Compute the Karl Pearson’s correlation Coefficient and comment on the results
C) Compute and interpret the Coefficient of Determination
d) Fit the Linear regression equation and comment on the results,
Time series notes 10/09/2024 19
e) What will be the sales income if advertisement expenditure is 50 millions
Time series notes 10/09/2024 20