Analysis of Sales and Customers

Analysis of Sales and Customers
Project A: Predicting Sales
To predict future sales of the supermarket, I built a multiple regression model whose
dependent variable was “Sales” from the historical data provided. First, I built scatter diagrams
and a correlation analysis to select the independent variables to incorporate in the model. From
the scatter plots, the variables that were found to have a strong linear relationship with the
variable sales were Wages $m, No. Staff, Adv.$’000, and Car Spare. The correlation matrix
indicated that the variables that had a strong relationship with sales were ‘Wages $m’, ‘No.
Staff’, ‘Adv.$’000’, ‘Competitors’, ‘SundayD’, ‘Mng-Exp’, and ‘Car Spares’. Therefore, I
choose to include these variables in the regression model for further analysis.
Using seven independent variables for the multiple regression model, R2 = 84.12% which
means that the joint variation in ‘Wages $m’, ‘No. Staff’, ‘Adv.$’000’, ‘Competitors’,
‘SundayD’, ‘Mng-Exp’, and ‘Car Spares’ accounts for 84.12% of the variation in “Sales”. I used
the F test to verify whether there was any substantial relationship between the dependent factor
and the set of the independent factors.
H0: b1 = b2 = . . . = b7 = 0 (there is no significant relationship between the dependent
variable and the independent variables)
H1: at least one b1≠ 0 (there exists a strong relationship between the dependent variable
and at least one of the independent variables)
From the regression output, ‘Wages $m’, ‘Adv.$’000’, ‘Competitors’, and ‘Mng-Exp’: p-
value = 0.000 < 0.05, Reject H0. These variables are all important in the model
For ‘SundayD’: p-value = 0.010 < 0.05, Reject H0. The variable is significant in the
model.