Locating New Pam and Susan Stores

Locating New Pam and Susan’s Stores

Stephanie Larsen

MGSC6200 50585 Information Analysis

Professor Anthi Tsouvali

July 5, 2015

Introduction

        The purpose of this case study is to analyze and identify where a new location of Pam and Susan’s Stores should be sited.  Given the data sets, we must recognize which site has the highest earning potentials and sales.  By using a multiple regression model and the historical data given, we will deduce which site will be best suited for the new location of Pam and Susan’s Stores.  

Data

        Several data types were used for this project analysis. Data was provided in a file that consisted of population, economic status, store size and sales from current Pam and Susan’s stores. The variables we were given are as follows:

1. Store size: Gross sq. ft. and selling sq. ft.
2. Competitive group : subjective judgment based scores (on a scale of 1 to 7)
3. Population: % of population who are black and/ Spanish-speaking
4. Family income(in $1000s) : 0-10, 10-14,14-20, 20-30, 30-50, 50-100, >100
5. Median yearly income
6. Median rent per month
7. Median home value
8. % of population who are homeowners.
9. % of population with no car.

1. % of population with 1 car.
2. % of population with TV.
3. % of population with washer.
4. % of population with dryer.
5. % of population with dishwasher.
6. % of population with air conditioner.
7. % of population with freezer.
8. % of population with a second home.
9. Education (in years) : 0-8, 9-11, 12, 12+
10. Total Population
11. Average family size

Results and Discussion

        The goal of this project is to predict where the new location of Pam and Susan’s stores should be located.  Since we have established a subjective judgement based score, we evaluated those by using a scatter plot to see if we could deduce any data (see Figure 1).  We can see from the comtypes in the data that only comtype 1, 2, and 7 would suffice to analyze this data.  Comtypes 1 and 2 have higher sales, and comtype 7 proves to have lower sales.  When we create dummy variables, these comtypes will likely be statistically significant for our findings.  To make use of these categories we created seven dummy variables (comtype1 = 1 if and only if a store has competitive type=1 and 0 otherwise.), although we can see that other than category 1 and 2 all the others are very similar and we decide to work with category 7 as it is the other end of the spectrum.  These were then were used in multiple regression models illustrated in Excel numbered Model 1-7.