Peoria's Best Coffee House
Spring 2015 Extra Credit Question
There is a maximum of 10 possible points.
NOTE: This question uses the Excel file titled Peoria's Best Coffee House Extra Credit Raw Data 2015.
The Backstory
Bob Roaster, the vice president for sales at Peoria's Best Coffee Houses (located in suburban Prescott Florida) is considering expanding this company's chain of coffee houses to Tallahassee. But before he can commit to opening one or two stores in the state capital, he needs to figure out what size stores he might open and how much could they make in revenue per week. He wanted to ask his friend Marvin Bloomy at PC (Prescott College) for help, but old Marvin just left on a 4-week trip to Alaska to see the "sights and to commune with bears" (whatever that means) and is out of touch. So you will have to help him out.
Use StatTools for this extra credit. You will be doing a variety of statistical tests to determine how these variables compare, and a multiple regression using weekly sales as the dependent variable and store size, college nearby and drive thru as the independent variables. Refer to the chats, textbook and/or PowerPoints on how to do and analyze this data and to use the results of the regression to predict the best type of coffeehouses Bob should build in Tallahassee and where they should be located.
Remember to add the variables by first using the Data Set Manager, and then calculate the following statistical tests. Be sure to explain them as you would in any professional reports. The report is to be presented to Bob Roaster from you as his professional statistician. (Lucky you!)
NOTE: The report, in the form of a mini report is to be in a single Word document submitted along with the Excel file that you used to calculate your answers. The date due is determined by the weekly/monthly calendar for your section.
1. Create a one variable summary of all 4 data elements and then analyze it.
a. What does it show you about the regression to follow? Be specific. Hint: Look at the mean versus the median and the skewness versus the kurtosis, what do they tell you about the 4 variables?
b. What other things can you learn about the data provided in the one variable summary?
2. Create histograms for weekly sales and size, (put a checkmark before both and they come out on the same output).
a. What do they show you about the data?
3. Create a box and whisker plot for both sales and size, what do they show you about the data?
a. Are there outliers in the data?
b. If so do you retain the outliers or delete them, and why?
4. Do a correlation on all 4 variables (StatTools/Summary Statistics/Correlation and Covariance/Correlation/Entries below the Diagonal).
a. Which of the variables has the greatest impact on sales?
b. Which of the variables has the least impact on sales?
c. What else can you determine from the correlation?
5. Create a multiple regression with the weekly sales as the dependent variable and the other 3 as the independent variables.
a. Discuss the r-square value for these results. How much of the variation has NOT been predicted by the regression?
i. What additional information would you want to have to increase the predictive validity of the data? Be specific.
b. Are the results sufficiently high to use for prediction?
i. Why or why not?
c. Is the "F" value significant or not?
i. What does this "p-value" mean?
6. Compare the residual statistics generated by the software - what do they tell you about the residuals in this case?
7. Predict what the weekly sales would be for the following information of 3 typical coffeehouses(using the coefficients from this sample) :
Table 1
Store 1 Store 2 Store 3
Store Size 750 955 1150
College yes yes no
Drive Thru no yes no
Estimated Weekly Sales
8. Based on this data what type of coffeehouses should Bob construct in Tallahassee and where should they be built? Be as specific as you want but answer the question.
NOTE: Attach your Excel file to this document that shows that you have completed this analysis of the data. Points will not be given unless this Excel file is attached as an attachment along with your answers.