Convenience Shopping

...

Last, to demonstrate which factor or combination of factors have impact on sales of convenience store, we construct both multiple and simple regression models.

[pic 16]

R-squares

sales on volume (model 1)

0.4219

sales on car washes (model 2)

0.0289

Sales on volume an car washes (model 3)

0.43

R2 is the percentage of the variation in the sample of y values that is explained by the regression equation. Larger R-squares means collected data are closer to the regression line. Therefore, we can conclude model 3 is the best model.

Both car washes sold and volume of gas sold have impact on sales of convenience store. The conclusion is plausive because nowadays people enjoy one-stop service, so when people go to refuel for their vehicle or wash their car, they will go shopping as well. [pic 17]

Moreover, to confirm the dependent and independent variables, we also did a reverse test for our data, we set volume of gallons as dependent variable and sales of convenience as independent variable. We draw the scatter plot and find out there’s nonlinear relationship between them.

Practical application

Now let’s turn our statistic results into real life.

Referring from second part, the correlation between volume of gas sold and sales of convenience store is strongest among other pairs of set. There are some reasons for the strong correlation between them: first, when people go to the gas station to refuel for their car, they might not care to spend more time going shopping in the convenience store; second, most convenience stores, such as Dillion’s and Kum&Go, issue membership card to their customers. When customers use the membership card to refuel, they will get discount, which is depend on the points on their card. After fueling their vehicles, the customers may want to raise their points, so they will get more discount next gas refueling.

In order to research whether the currency rate and unit of gas will affect our result, we decide to change our currency to Canada currency and transfer the unit of gas from gallon to liter. By constructing a new regression model, we find out that the R-square keep the same. In other words, the currency rate and unit of gas do not have influence on our outcome. [pic 18]

R-squares

Intercept

slope

sales on volume (model 1)

0.4219

1263.03854

0.09316

sales on the number of car washes (model 2)

0.0289

2327.41712

0.47953

Sales on volume and on the number of car washes (model 3)

0.4300

1112.1759

0.31503

0.1166