Vintage Restaurant Case Study

899 Words2 Pages

Karen Payne, owner of the Vintage restaurant in Ft. Myers, Florida is looking to predict future food and beverages sales with their current sale history. The Vintage restaurant, a high-quality establishment, has just completed its third year in operation and provided us with their monthly food and beverage sales. Upon initial analyzation of our data, we could easily tell that there was a seasonality trend in the data provided to us. Some additional statistics about the data follow:
• September was the lowest month of sales (each year)
• December was the highest grossing month of sales (each year)
• From Sept. to January, sales would grow (each year)
• From January to Sept., sales would decline (each year)
With this in mind, I then began …show more content…

With that coefficient irrelevant, the numbers of the above coefficients are slightly different to provide an adjusted forecasting model.
According to the data above, forecasted sales (in thousands) equals 223.667 + the correlating coefficient depending upon the month of the forecast. Furthermore, this indicates that before the business was opened, they were earning $223.667(thousand). Obviously, did not happen so we can disregard this value due to extrapolation.
With this regression we were also provided an R squared value of .948 or 94.8% meaning that 94.8% of the variability in food and beverage sales can be explained by the month coefficient.
With two indicators now showing that the model with a trend is providing more accurate results (graph and the R-squared value), I felt slightly more confident in using that forecasting model. To be completely sure however, I decided to verify with one more test: the mean squared error. The mean squared error measures the average of the squares of errors, - the lower the MSE, the better. Overall, the mean squared error for our data with a trend was 12.6 indicating that the model closely resembles the data that we were provided. For our linear regression model without a trend, the MSE was calculated at 111.96 indicating that the model is not as closely related to

Open Document