This multiple regression project which is based on Pam and Susan’s stores is designed to help us analyze the most efficient and effective way of determining how to predict the sales potential at new store locations. After being given several variables which are based from the census of the trading zone from the current 250 stores, it is my job as an outside consultant to compile this data and determine what type of growth potential will be developed from the new location sites.

The information that was provide for analysis consisted of sales per stores (which is the y variable), along with 33 x variables such as ethnicity, income, population, home owners, renters, etc. There were several observations done to analyze the data which consisted of using scatter plots, running correlation and residual models. By performing these models I was able to then identify the most significant of the remaining variables and then remove the variables that are insignificant.

The first steps of my analysis consisted of me creating a scatter plot for sales vs comtype. By doing this I was able to determine the variances between the different comtypes and identify which comtypes were significant and insignificant. It is clear based on the scatter plot below that the comtypes 1 and 2 have significantly higher sales than comtype 7 while the other comtypes are within the same range. Based on this information I know that comtypes 1, 2, and 7 are significant variables.