Business Analytics

This document discusses various topics related to Business Analytics including visualization tools, regression analysis, and classification. It provides insights on the CommBank Retail Business Insight Report FY2018 and the application of regression analysis in determining consumer behavior trends. It also explains the concepts of classification and prediction in analytics.

BUSINESS ANALYTICS

BUSINESS ANALYTICS
QUESTION 1
Part i
The CommBank Retail Business Insight Report FY2018 appropriately utilizes visualization tools
for the presentation of the information. The visualizations used in the report are both of high
quality and relevant for the type of information being communicated. The arrangement (layout)
of the report consists of separate layers for the written information and the visualized
information. That is, the written section and its visualization are side by side. Although this may
look convenient, a more downward layout would be better for transition in reading the report.
The information provided by the report is both specific and comparative, which gives the readers
both a deeper and wider view of the subject of the report.
Part ii
1. A majority of the retailers are innovation active.
2. The multichannel retailers are the most innovation active among the categories of retailers.
3. The biggest innovation drive for retailers is the improvement of efficiency and production.
4. The key perceived benefit of innovation is improved market position with lack of time
dedicated to innovation being the biggest challenge to innovation.
5. Sales and marketing, and online presence lead in the innovation areas where retailers invest.
6. The average return on investment on innovation in the retail industry is almost double the
investment.
The information above is relevant in determining whether to invest in innovation or not. Through
this information, a retailer will be able to approach investment on innovation based on its retail
2

BUSINESS ANALYTICS
category and with insights on the potential return on investment and market trend. Thereby,
enabling the retailer to make an optimal decision on investing on innovation.
Part iii
The CommBank Retail Business Insight Report FY2018 is a report that presents findings on the
retail business trend with regards to innovations. The report is focused on the retailers in
Australia for both regional and national retail businesses. This report collected information from
a sample of 2473 individuals in the retail business across Australia as well as from 16 interviews.
It derives information regarding the uptake, nature, expectation and challenges of innovation in
the retail industry in Australia.
Part iv
The report can be improved by firstly adopting a downward layout where written information is
followed by it visualization below. This will enhance the transition in reading of the report. The
report can also provide more summary information regarding the analysis techniques applied.
This will improve the credibility of the report by opening it to counter-checks.
QUESTION 2
Part i
An example of the application of regression analysis is in determination of consumer behavior
trends. Understanding consumers remains an important part of business to date. By using
3

BUSINESS ANALYTICS
regression analysis, the factors that affect a consumer’s decision to purchase or quantity of
purchase can be determined. These factors can then be targeted so as to ensure purchase or
increase in purchase by a consumer.
Part ii
The table below, Table 1: Height-Weight Data, represents the information on the height and
weight of 10 students:
Table 1: Height-Weight Data
Student Height.ft. Weight.Kgs
1 6.9 56
2 5.1 58
3 7.5 58
4 4.4 57
5 5.3 55
6 6.3 54
7 2.8 55
8 4.8 57
9 5.2 56
10 5.1 54
Part iii
The plot in Figure 1: Scatterplot for Height-Weight Data below represents a scatterplot of the
height and weight variables in Table 1: Height-Weight Data above. From the plot, we observe
that there is no linear trend among the variables. Hence, there is no significant linear relationship
between the height and weight variables.
4

BUSINESS ANALYTICS
2 3 4 5 6 7 8
52
53
54
55
56
57
58
59
Height-Weight Scatterplot
Height.ft
Weight.Kgs
Figure 1: Scatterplot for Height-Weight Data
Part iv
The regression equation for the height and weight variables is as follows:
Letting Weight be X and Heights be Y, then the means would be 56 and 5.34 respectively with
totals as 560 and 53.4 respectively as well. Table 2: Parameter Estimation below shows the
computations for the parameter estimation.
Table 2: Parameter Estimation
Xi –
Xbar
Yi -Ybar (Xi -
Xbar)^2
(Xi-Xbar)(Yi-Ybar) (Yi -Ybar)^2
0 1.56 0 0 2.4336
2 -0.24 4 -0.48 0.0576
2 2.16 4 4.32 4.6656
1 -0.94 1 -0.94 0.8836
-1 -0.04 1 0.04 0.0016
-2 0.96 4 -1.92 0.9216
-1 -2.54 1 2.54 6.4516
1 -0.54 1 -0.54 0.2916
0 -0.14 0 0 0.0196
5

BUSINESS ANALYTICS
-2 -0.24 4 0.48 0.0576
20 3.5 15.784 Totals
β1=∑ ( Xi− Xbar)(Yi−Ybar )
∑ ( Xi− Xbar)2 =3.5
20 =0.175
β0= ∑ Y −β1 ∑ X
n = 560−0.175 x 53.4
10 =4.6655
Thus, equation is given as:
Y =4.6655+0.175 X
Part v
The R2 is given as follows:
R= ∑ ( Xi− Xbar)(Yi−Ybar )
√∑ ( Xi−Xbar )2
∑(Yi−Ybar )2 = 3.5
√20+15.785 = 3.5
17.77 =0.1970
R2=0.19702=0.0388
The value of the R2 given above indicates that the model fitness is equal to 3.88%, hence not a fit
model.
Part vi
Table 3: Excel Regression Analysis Summary below represents the output of the regression
analysis. From the table, the regression equation is given as below. We observe that the value of
the intercept differs from the calculated value from Table 2: Parameter Estimation, however the
6

BUSINESS ANALYTICS
value of the coefficient for x remains the same. We also observe that the value of the R2 remains
the same as above.
Y =−4.46+ 0.175 X
Table 3: Excel Regression Analysis Summary
QUESTION 3
Part i
Classification refers to the process of categorizing a group of objects or individuals depending on
similarities in their attributes while prediction is the process of determining the value of an
attribute of an object or individual at future time.
Part ii
1. Neural Networks
7

BUSINESS ANALYTICS
2. Clustering
Part iii
yi=w 5 { 1
1+ e− (b 1+i 1 w 1+i 2 w 3 ) }+ w 6 { 1
1+e− ( b2 +i 1 w2 +i 2 w 4 ) }
In terms of classification, the final output of the neural network is subjected to a cut-off value to
determine where an observation of inputs will be categorized. Consider a case of two categories
“Yes” and “No”, the cut-off can be set as 0.5 such that if the value of the final output of a neural
network is above 0.5, then the observation of inputs is categorized as “Yes” while if the value of
the final output is below 0.5, then the observation is categorized as “No”.
Part iv
1. Consider a bank that is interested in understanding its clients depending on their loan
repayment rates. Clustering can be applied in this case to group together clients with similar
loan repayment rate after which the groups can be observed for characteristics.
2. Consider a portfolio manager that is interested in determining which portfolios have similar
performance margins at the stock market. Through clustering, the portfolio manager can
generate groups of portfolios with similar performance margins.
3. Consider a juice producing company that supplies to retailers and is interested in finding out
which retailers order same amounts. By using clustering, the retailers can be classified into
groups with similar order amounts.
8

BUSINESS ANALYTICS
9