CommBank Retail Business Insights Report FYI18

This report provides insights into the retail sector of the business for the financial year 2018. It includes information on the number of businesses surveyed, innovation performance, investment areas, and returns from innovation.

MITS6002 Business Analytics
Name. email. Date. Professor.
Q1. The CommBank Retail Business Insights Report FYI18.
i. According to this report, we notice that the formatting of the insights is appealing. The use of high
capturing headers in the pdf formatted paper to represent the key reporting issues makes it captivating to
read the whole paper. As depicted by the overview, we already note the location of the business and thus
one can be able to get the scope just from the overview of the report. Use of auto-capturing and mind-
blowing visual charts and images, graphs make it so much easier to understand the textual part. This
strategy helps one to keenly understand the numbers in a deeper analogy. Use of maps and pictures even
the more makes the report glowing. We can see that the information provided is for the financial year of
2018 specifying the retail sector of the business.
ii. According to CommBank report [1], we take a note of the 2473 business organizations that were under
the survey. The sampling size used for analysis is 262 retail business of which the study indicates that
about 87% are found in the brackets of either being Active Innovators or Improvers. Further looking at the
innovation performance section, we notice that 71% retailers in the multichannel section are all in the
Innovation Active Zone.
Following this trait, we get to make comparisons and contrasts between the NationalFY18 and ReatilFy18
in terms of the investment areas [1]. Such a comparison can be well depicted by also employing the use of
some nice inferential statistics such as hypothesis testing. This will well get the significance of each
variable under statistical investigation. The figures show that 48% investment is done in the sales and
marketing area and 55% investment is done using on the websites or digital presence [1]. Here we find a
marketing strategy that can best be optimal for a new retailer to delve into. There is a revelation that
technology plays handy for investment from the chart comparison in the section “Investment in
innovation”.
Finally, the most crucial part comes in which is the number of returns from innovation. Figures depict that
about 80% of the retailers do invest with a high expectation of ROI in a twelve-month period [1]. This
helps a new investor be objectivial in the ROI returns and not to make over ambitious plans during the
investment periods. We can note that the average investment amount for the retail investment is $101,000
where by there is an average additional earning of $198,000 and nice average return of 1.96 [1].
iii. Abstract
In this research paper, we have been looking at the details in retail business especially in the
contribution towards the banking sector of the economy through innovations. The main idea in the
methodology employed is the use of descriptive analytics with a much delving into the data
visualization. As seen in the report, some of the retailers are seen left behind in finding the financial
and other relevant benefits arising from the innovations sector [6]. The dynamic nature of the
factors influencing the innovations progress have been critically analyzed to find the way different
industries will adapt to these factors. There has been in fact a concentration on retail traders
making innovations tending to only a few patterns without discovering other key player
1

opportunities. As seen by the analytics from CommBank reporters, one can get the easiest means of
doing advertising basing on the fact that many businesses have to market their innovative strategies
in order to attain high profit. On this note, we can see the ROI [6] is highly attributed to the
conjugation of high-end factors of the innovation process. This therefore means that the budget
making process will rely on activity that pump returns on a short-term basis. In business, we all
understand that the underlying policy it to cut on cost as we maximize the saving. Cost-benefit
analysis therefore explodes the analytics process.
Key words: Data, CommBank, Innovation, ROI, Analytics.
iv. In this insights report, reflexive actions have to be taken in order to suggest a design of subjective
methods to be done. Use of regression modelling to come up with comparison models between
RetailFY18 and NationFY18 would be deemed a high-end process to describe different patterns in the
insight report. A recommendation to check the significant variables by setting and testing different
hypothesis and claims would be another measure of improving the insight report. Finally, we can also use
the classifications models to build easy predictable models that would build the analysis process into
meaningful insights.
Q2. Regression as a commonly used technique to finding relationships amongst variables.
i. Regression analysis can be used in the stock market. It can be used to predict stock based on the times
variable and we can add Dummies as variables for the interval months say 11 dummies for month wise
data or 3 Dummies for quarterly data. This kind of regression is called a time series regression analysis,
which can be used to supplement insights from other time series methods.
ii.
Name Height Weight
Jenny 58 115
DanJonso
n 59 115
Amelia 62 145
Hillary 63 120
Ava 65 133
Chloe 67 135
Lucas 68 142
2

Evie 68 140
Trinta 71 146
Noah 72 157
iii. Scatter plot
.
58 60 62 64 66 68 70 72
120 130 140 150
Relationship Between Weight and Height
Height in Inches
Weight in Pounds
Figure 1. Scatter plot depicting relationship between height and weight.
As seen by the above plot, we can comment that there exists a linear relationship between height and
weight. The relation is positive in that as weight increases the weight too increases. This implies a strong
positive correlation. A correlation matrix can also be constructed to determine the amount of relationship
between the two variables.
iv. We use the method of least squares to fit the line. Let the independent variable be height while the
dependent is weight.
Finding the regression Line using Least Squares Method.
3

Height (x)
Weight
(y) X2 Y2 X*y
58 115 3364 13225 6670
59 115 3481 13225 6785
62 145 3844 21025 8990
63 120 3969 14400 7560 N=10
65 133 4225 17689 8645
67 135 4489 18225 9045
68 142 4624 20164 9656
68 140 4624 19600 9520
71 146 5041 21316 10366
72 157 5184 24649 11304
Totals 653 1348 42845 183518 88541
calculating the Slope M
2.5311122
Calculating the Intercept
B;
-30.4816267
Thus the regression equation becomes
Weight = 2.5311122* height -30.4816267
v. Computing the R2 value and commenting
vi. on the goodness of fit.
Calculatin
g R-
Squared..
Height (x) Weight (y) xbar ybar x-xbar y-ybar (x-
xbar)*(y
(x-
xbar)^2
(y-
ybar)^2
4

-ybar)
58 115 65.3 134.8 -7.3 -19.8 144.54 53.29 392.04
59 115 65.3 134.8 -6.3 -19.8 124.74 39.69 392.04
62 145 65.3 134.8 -3.3 10.2 -33.66 10.89 104.04
63 120 65.3 134.8 -2.3 -14.8 34.04 5.29 219.04
65 133 65.3 134.8 -0.3 -1.8 0.54 0.09 3.24
67 135 65.3 134.8 1.7 0.2 0.34 2.89 0.04
68 142 65.3 134.8 2.7 7.2 19.44 7.29 51.84
68 140 65.3 134.8 2.7 5.2 14.04 7.29 27.04
71 146 65.3 134.8 5.7 11.2 63.84 32.49 125.44
72 157 65.3 134.8 6.7 22.2 148.74 44.89 492.84
Total
s 653 1348 653 1348
2.84E-
14 -1.1E-13 516.6 204.1 1807.6
Numerator 516.6
Denominato
r 607.397
r
0.85051
5
Rsquared
0.72337
5
From the above Rsquared, we can see that about 72.33% of Weight of family member can be accounted
for by their height. This is thus a good model.
vii. Using Rstudio software, we can use the lm function to get the regression summary.
5

The greens highlight on the above table are the solution to iv and v.
We can see that the method of least squares has given the exact solution as the one generated by the
software.
Q3. Classification and Regression as used in business analytics.
i.
When we talk about classification, we refer to the idea of finding the classes of variable or sample
variables in a dataset. This involves grouping a particular data into the categories where the basis is a
particular training set.
Looking at prediction, this is a statistical tool that involves making prediction especially over a missing
set of items in a dataset. Here, we require the different methods of classification or regression in order to
come up with prediction models of missing or unknown items.
ii. Types of classification methods or algorithms include:
a) Linear classifies such as linear regression, logistic regression, Naïve Bayes Classifiers.
b) Decision Trees.
c) Neural Networks.
d) Random Forest.
e) Nearest Neighbors.
6

f) Support Vector Machines.
g) Just to mention but a few.
iii. The type of network shown is a feedforward neural type of network. Its equation can thus be written in the
form of.
F(W, X) =∅ (∑
1
2
Xi∗Wi)
The inputs are 2 where here they have been represented by Xi instead of Ii.
Neural network mimics the human brain in their operation modes. They represent and algorithm used to
recognize different pattern sets in a data. Specifically, they are created in form of nodes and used for deep
learning. What a node does is combination of the input from the data with a set of coefficients and their
respective weights. Then summation of the input-weight and then passed to another node which then
determines the extent of the signal. This act of passing spectrums from one node to another of the same
type leads to classification. The most widely being the binary classification that groups in terms of true
false.
iv. How different classification methods are applied to business analytics.
a) We can use Naïve Bayes classifiers to determine the credit card detection in banking and finance
sectors.
b) We can also employ logistic regression to make come up with predictive models for sales of goods
and services in a company.
c) K-nearest neighbors can be used by a sales and marketing agent to determine the shortest routes in
combination with the decision trees to help them find the situation of company branches.
References.
[1]Commbank.com.au,2019.[Online].Available:
https://www.commbank.com.au/content/dam/commbank/business/pds/retail-business-insights-report-fy18.pdf.
[Accessed: 10- Jun- 2019].
[2]"Regression Analysis by Example, 5th Edition", Wiley.com, 2019. [Online]. Available: https://www.wiley.com/en-
us/Regression+Analysis+by+Example%2C+5th+Edition-p-9780470905845. [Accessed: 07- Jun- 2019].
7

[3]E. Steyerberg, "FRANK E. HARRELL, Jr., Regression Modeling Strategies: With Applications, to Linear Models,
Logistic and Ordinal Regression, and Survival Analysis, 2nd ed. Heidelberg: Springer.", Biometrics, vol. 72, no. 3, pp.
1006-1007, 2016. Available: 10.1111/biom.12569.
[4]"Excel 2013 Statistical Analysis #31: Create Discrete Probability Distribution, Calculate Mean and SD", YouTube,
2019. [Online]. Available: https://youtu.be/IQDgmmYXrAI. [Accessed: 04- Jun- 2019].
[5]"Big data analytics and business analytics", Taylor & Francis, 2019. [Online]. Available:
https://orsociety.tandfonline.com/doi/full/10.1080/23270012.2015.1020891. [Accessed: 07- Jun- 2019].
[6]N. Ardjmandpour, C. Pain, J. Singer, J. Saunders, E. Aristodemou and J. Carter, "Artificial neural network forward
modelling and inversion of electrokinetic logging data", Geophysical Prospecting, vol. 59, no. 4, pp. 721-748, 2011.
Available: 10.1111/j.1365-2478.2010.00935.x.
8