You are on page 1of 5

a.

Total Revenue Distribution:


The following plot shows distribution of total revenue:

From this plot we can infer the following:

 Data Range (Rev_Total): 0-94


 Though the data extends from 0-94, most of the data is concentrated in the region
between 0 and 20 (Rev_Total) and it is normally distributed within this particular range

b. Model:
Rev_Total = bt0 + bt1 * Bal_Total + bt2 * Offer + bt3 * CARD + bt4 * LOAN + bt5 *
INSUR + bt6 * Check
c. Model Outputs:
d. Conclusions:
Based on the results we can explain the impact of each and every independent variable.
Overall the model is significant and the R-Square is 0.11.
Rev_Totalcustomer=0.55143+0.000053Bal_Totalcustomer+0.2992Offer1cu
stomer+1.24CARD1customer+2.27LOAN1customer+0.3798INSURcustomer+0.
00549Checkcustomer + €customer
As the model is significant, we have to reject the null model and need to explain the effects of the other
determinants.
All the variables have positive betas.
Bal_Total
The revenue will increase by 0.000052961units for every unit increase of account balance of
the customer if all other variables remains constant.
Offer1
The revenue will increase by 0.299130817 units if the customer has received a special
promotional offer in the previous one-month period provided all other variables remains
constant.
Card1
The revenue will increase by 1.241809255 units if the customer has an active credit card
account than a low or zero active credit card account provided all other variables remains
constant.

Loan1
The revenue will increase by 2.273504438 units if the customer has an active loan account
than a low or zero active loan account provided all other variables remains constant.

Insur1
The revenue will increase by 0.379775009 units if the customer has an active insurance
account than a low or zero active insurance account provided all other variables remains
constant.
Check
This variable is not significant.

e. Predictive Ability of the model:


The model can explain only around 11 percent of the y variable (Revenue) using the current
set of predictors.

f. Limitations of the model:


The limitations of the model are listed below
 The data is not panel data
 We don’t know the customer acquisition and retention cost.
 The loan interest rate might change over the time period and hence should be one of
the predictor variables.
 There data is not normalized.

SAS Code:
data x1; set bank_rev.bank_rev;
run;

proc contents data=x1;


run;

proc sgplot data=x1;


histogram Rev_Total;
density Rev_Total;
run;

data x2; set x1;


sno=_n_;
run;
data x3; set x2;
if sno le 6500;
run;

data x4; set x2;


if sno gt 6500;
run;

proc glm data=x3;


model Rev_Total=Bal_Total Offer CARD LOAN INSUR Check /solution;
store out=params1;
run;

proc plm source=params1;


score data=x4 out=model2;
run;

You might also like