Business Analysis

This document discusses the features of Insight Report FY18, Modelling of Regression Equation, Classification and Prediction, Analysis of Neural Network, and Applications of Clustering in Business Analytics.

Running head: BUSINESS ANALYSIS 1
Business Analysis
Student Name
Professor’s Name
University Name
Date

BUSINESS ANALYSIS 2
Business Analysis
Question 1: Case Study: Insights Report FY18
I. Features of the insight report.
a) Visualization
Visualization is a technique used in statistics to summarize data in graph or
pictorial form so that meaningful conclusions can be drawn easily and quickly.
The insights report uses graphs such as histograms and bar charts to summarize
important aspects of the data (Levie, 2012). A reader can easily and quickly
identify important facts and draw conclusion without the necessity of reading
through the entire text. The use of text with different font types and style has also
been applied as a way of making the insight report more attractive to the reader.
b) Presentability
The information provided by the insight report is presented in different formats.
For example, there is the use of graphs and pictures. The graphs help the reader
understand the data and make important conclusions from the data easily. The
picture helps the reader know what the report is all about even before reading the
report. A summary about all that has been presented in the insight is also provided
at the beginning of the insight report so that a reader can quickly know the key
information the insight report provides for the reader. The language is used is
fairly simple of any reader to understand since it’s a simple plane language that is
fairly not too much statistical (Lock, 2013).
c) Information Provided

BUSINESS ANALYSIS 3
The facts presented by the report is detailed and has been delivered in a summary
form. The title provides clear description of the information contained in the
insight report. Also, the insights information uses short paragraphs that only detail
the key and necessary information required determine whether the actual
objectives of the paper have been met. There is also a logical flow of information
in the insight report for easy readability and understanding.
II. Information Retrieved from the Insights Report and its application in decision
modelling
The purpose of the insight report is to show how the retail industry in Australian is
putting into action good innovative techniques to curb the existing competition in the
industry while maintain customer experience and improve the efficiency and growth. As
a result, the following facts can be acquired from the insight report.
 The current status of the retail industry in Australia based on the information
provided by the CommBank. This information is used as a bench mark upon
which research studies indicate whether there has been an improvement or not.
 Various innovative techniques that have been employed by the Australian retail
industry to grow the efficiency, the market and make use of available
opportunities while maintaining customer experience. The information about the
innovative techniques can help determine if all the best innovative techniques
have been exploited, if there are more that can be embraced and the impacts the
innovative techniques have on the retail sector.

BUSINESS ANALYSIS 4
 The performance of the retail sectors that have adopted the innovative techniques
or are improvers of innovation. It can help to make decision on whether to
embrace or to reject innovative techniques.
 The investments of the retail industry on innovative techniques, the returns on
investments and the amount of time they have to wait to achieve the returns of
investments. The information can also help make the decision of whether to
embrace or reject innovative techniques.
III. Abstract Summary of the Insights Report
There exists a lot of competition in the Australian retail sector. The sector is toiling
tirelessly to limit the effects of the competition so as to maintain the growth while
ensuring the efficiency and the customer experience is improved. The Australian retail
sector strives to achieve the same through adoption of innovative mindsets that maximize
the available opportunities. An analysis on the data provided by CommBank indicate that
87% of the sector activities are either innovative or improvers of innovation and 71% of
these activities are undertaken by multichannel retailers. To achieve high levels of
innovation different techniques of innovation are employed. For example, 48% of the
retails have put their money on innovative techniques of sales and marketing while 55%
bank on digital technology such as websites. 80% of the enterprises that have put their
money on technology expect to have attained their returns on the investment within a
year.

BUSINESS ANALYSIS 5
IV. Recommendations.
The quality of the insight report is good, it is well organized, uses good visualization
and presentation methods. However, the quality of the report can further be increased if
the following key factors were to considered.
 Include varied pictorial and graphical displays such as pie charts, heat maps,
boxplots etc. Currently, the insight report only utilizes histograms and bar charts
and considering its length, monotony would result into a reader not taking keen
interest into some of the histograms and the bar plots.
 The length of the insight report can also be reduced. Most of the words presented
in text form can be converted into images delivering the same information. This
way a reader can not get bored and can draw meaningful information quickly.
 The insight report could opt to use simple non-statistical language that can easily
be understood by all individuals even those who are not technical enough to
understand statistical terms.
 The choice colors for displaying the different information could be selected in
such a way that they are more appealing to the readers eye. The colors used
currently are quite dull and could inflict boredom into the reader.
Question 2: Modelling of Regression Equation
I. Example of fields of application of regression analysis
 Forecasting and prediction: regression is used to predict the probability of an
event happening based on various other factors. For example, it is used to predict
the possibility of a team winning or losing a football match or price of products is
affected by demand and purchase patterns of customers (Shao, 2010).

BUSINESS ANALYSIS 6
 Optimization: result of regression can be used to optimize process or activities in
an organization. For example, it can be used to optimize manufacturing time in an
industry or the waiting time in a hospital (Rumsey, 2015).
II. Regression Data Collected from Colleagues.
That table below illustrates the heights of ten of my colleagues and the related weights.
III. Scatterplot of the heights and weights
The scatter plot for the association between height and the related weights.
The Plot depict that the relation between weight and height is positive and linear.

BUSINESS ANALYSIS 7
IV. Regression Equation
The equation is:
y=c+ mx
The data used in computation is shown in the table.
The value of the intercept c indicate that the value of dependent variable which is
weight when the independent variable height is null. The value of the intercept is given
by:
c= ( Σ y ) ( Σ x2 ) −( Σ x)( Σ xy )
n ( Σ x2 )−¿ ¿
c= ( 604 ) ( 265651 ) −(1621)(98908)
10 ( 265651 )−¿ ¿
c= 123336
28869
c=4.27
The slope represents coefficient of the height indicating the magnitude with which
the height affects the weight.
m= n ( Σx y ) −( Σ x)( Σ y )
n ( Σ x2 ) −¿ ¿

BUSINESS ANALYSIS 8
m= ( 10 ) ( 98908 )−(1621)( 604)
10 ( 265651 )−¿ ¿
m= 9996
28869
m=0.346
The regression equation:
y=0.346 x +4.27
The regression equation shows that the values of weight will be 4.27kg if height is zero
and 4.616kg if height is 1. This is an indication of positive linear relationship.
V. The coefficient of determination and the goodness of the fit.
r =n ( Σx y )−(Σ x )(Σ y )
√¿ ¿ ¿
r = ( 10 ) ( 98908 ) −(1621)(604 )
√ [ 10 ( 265651 ) − ( 1621 ) 2 ] [10 ( 36926 ) − ( 604 ) 2 ]
r = 9996
11326.69
r =0.8825
The coefficient of determination:
r2=0.88252
r2=0.78
The r-square value indicates 78% of the variability in the relationship between the
variables can be described by the model.

BUSINESS ANALYSIS 9
VI. Using excel to compute regression for comparison purposes.
Excel is used to compare the results. The regression model output is:
The equation is:
y=0.346 x +4.27
The coefficient of determination 0.78 and therefore the result obtained through manual
calculation are similar to the ones obtained in excel.
Question 3: Classification and Prediction
The two are statistical techniques that are used in prediction and forecasting of trends and
patterns in sets of data (Hinton, 2014). They are used in extraction and retrieval of models which
are used to examine and describe classes of important information that are of great importance
(Freund, 2014).
I. Distinguishing features of Prediction and Classification
The differences between classification and prediction are:
 In classification the classifier strictly classifies categorical variables while in
prediction the predictors strictly examine continuous variables.

BUSINESS ANALYSIS 10
 The accuracy in classification relies solely on correct and accurate definition of
class labels while the accuracy of prediction entirely depends on the how
accurately the predictor can predict the predicate attribute value with no errors.
 In classification identification is performed on the categories of data to which
there exists a possibility of existence of new observations and is based on accurate
selection of a training data set which contains the observations of the known
categories. On the other hand, missing numeric data are determined for the
observations in prediction.
II. Classification Methods
The most common methods of classification that are applied in statistics and data
mining for forecasting and prediction purposes are:
 K-Nearest Neighbors Method: This is a method of classification where the
training dataset is broken down into groups comprising of k number of
observations using the measure of Euclidean distance. The method determines the
distinct similar features in the nearest “neighbors.” The neighbors with similar
features are placed in the same group which are then used to define categories to
the members contained in validation data set (Linoff, 2011). `
 Neural Network Method: This method of classification utilizes a model that
works in a similar manner as the human brain. It comprises of neural networks
whose purpose is to process individual records at any given instance of time and
compare the classifications of the records arbitrarily with sets of well-known
record classifications. A feedback system exists such that in the case where errors
occur during the first classification, they are fed back into the neural network and

BUSINESS ANALYSIS 11
used to modify the algorithm of the network. Different iterations are performed
until the desired level of accuracy is reached (Fowler, 2009).
 Naïve Bayes Method: It is a method of classification where the training data set
is scanned to determine the records of the dataset where the predictors are the
same. The group that occurs frequently during the classification process is
assigned all the observations with similar predictors. If more predictors for new
observations are determined to have the same features as the ones placed in group,
then these new observations are placed to the group.
 Logistic Regression: It a method of regression that is different from the normal
regression method and is used when the output variable is dichotomous in that it
only has the possibility of taking two discrete values like yes or no. The logistic
regression model is used to examine the probability of the occurrence of the
output variable or the response variable.
 Decision Trees Classification Method: This induction method of classification
utilizes flow diagrams that closely resemble tree-like structures. The non-leaf
nodes of the tree structure indicate the test that is conducted in the variable
attributes while the leaf nodes represent the labels of the classes. The branches of
the tree structures indicate the solution or the result of the test conducted on the
variable attributes.
 Discriminant Analysis: This method of classification uses various sets of linear
equations and functions to predict new classes of observations with an unknown
class. The linear functions contain the predictors that are used in the prediction of
the new observations..

BUSINESS ANALYSIS 12
III. Analysis of Neural Network
Algebraic Equation of Neural Network
We are provided with a neural network and we are required to determine the
algebraic equation that best fits the neural network.
The algebraic equation y1 is defined by input values i1, i2, and the weights w is given by:
h1=σ (i1 w1 +i2 w3 +b1 )
h2 =σ ( i1 w2+ i2 w4 +b2 )
Solving the two simultaneous linear equations above and substituting the results in to
the equation below yields.
y1=σ ( h1 w5 +h2 w6 )
This is the equation of the neural network.
Principle of the Neural Network
The neural network works in a similar manner as the human brain. It comprises of
neurons placed in layer like structure. The input layer and contains the records used
as the inputs to the second layer known as the hidden layer. The output layer is made

BUSINESS ANALYSIS 13
up of the node of each class. The layers are associated with weights where an
algorithmic function is used to add the weights and map the results to the output.
The neural network process individual records at any given instance of time and
performs an arbitrary comparison of classifications of the records with sets of well-
known classifications. A feedback system exists such that in the case where errors
occur during the first classification, they are fed back and used to adjust the algorithm
of the network. Different iterations are performed until the desired level of accuracy
is reached.
IV. Applications of Clustering in business Analytics.
Clustering is an unsupervised technique of learning used in statistics and data mining
with an aim of grouping abstract objects in datasets into classes so that the objects with
similar distinct features belong to the same group or class while those with different
characteristics and features belong to other groups (Foster, Barkus &Yavorsky, 2016).
Examples of areas where classification is applied in business analytics are:
 Segmentation of product lines: This is the common use of clustering that is
majorly applied in shops and supermarkets. The products are clustered based on
their type, physical appearances, and other common characteristics such as
weight, size, flavor, brand and quality. Clustering of products helps customers to
easily identify the products based on their choice and preferences. Also, it helps
the marketing team easily identify the right prices for various products in the
selling premises (Selvanathan & Keller, 2017).

BUSINESS ANALYSIS 14
 Market research: Clustering helps in the marketing teams in defining their
market territories based on the most selling product in a given territory and the
choice and preference of people in that territory (Rugg & Petre, 2017). Individuals
in a given territory are grouped based on their demand of products, their purchase
patterns and their preferred mode of product presentation and delivery. This way,
the marketing team can easily focus a given product for a given population of
individuals thereby increasing sales.
 Anomaly and defect detection: products are clustered and placed in different
groups so that those products which meet a certain threshold criterion are paced in
one cluster and used as a bench mark upon which other products in the
manufacturing line are compared. Those products which do not reach the
threshold criterion are considered defective and not put in the bunch to be sold.
Clustering also help in the determination of fraudulent transactions. Transactions
meeting the set standards are clustered together and those that do meet this
criterion are considered as fraudulent (Newbold, Carlson, & Thorne, 2013).

BUSINESS ANALYSIS 15
References
Foster, J. J., Barkus, E., & Yavorsky, C. (2016). Understanding and using advanced statistics
(2nd ed). London: SAGE.
Fowler, F. (2009). Survey research methods. Thousand Oaks, Calif.: Sage.
Freund, J. E. (2014). Modern elementary statistics (12th ed.). Boston: Pearson.
Hinton, P. R. (2014). Statistics explained (3rd ed.). London: Routledge, Taylor & Francis Group.
Levie, R. D. (2012). Advanced Excel for scientific data analysis (2nd ed). New York, NY: Oxford
University Press.
Linoff, G. (2011). Data analysis using SQL and Excel. Indianapolis, Ind.: Wiley Pub.
Lock, R. (2013). Statistics: Unlocking the power of data. Wiley.
Newbold, P., Carlson, W. & Thorne, B. (2013). Statistics for business and economics. Harlow,
Essex: Pearson Education.
Rugg, G., & Petre, M. (2017). A gentle guide to research methods. Maidenhead: Open
University Press.
Rumsey, D. (2015). Intermediate statistics for dummies (1st ed). Hoboken, N.J.: Wiley.
Selvanathan, E. A., & Keller, G. (2017). Business statistics abridged (7th ed). South Melbourne,
Victoria: Cengage Learning.
Shao, J. (2010). Mathematical statistics (2nd ed). New York: Springer.