QNT2020
FOUNDATIONS OF PREDICTIVE ANALYTICS & DECISION
MODELING
Project on
Predictive
Models
Understanding Franchise
Performance
u Your boss wants your help to understand what drives the performance
of stores of the Subway franchise.
u Your boss helps you acquire data from last year’s sales and the
demographic factors of each store:
u SALES
u SQFT
u INVENTORY
u ADVERTISING
u FAMILIES
u STORES
u The EXCEL data file (Franchises.xlxs) has a data dictionary describing
the meaning of the variables.
Using Simple Linear Regression
Models to Understand Franchise
Performance
u Use the dataset in Franchises.xlxs.
u Build a model to establish which individual factors affect sales in our
franchises and how?
u Can we create a model to help design a store and achieve a
certain level of sales?
u Which factors are critical to the design of the store: area (sq ft),
inventory, advertising spend, size of the sales district, number of
competing stores?
Using SLR Analysis to Understand
Franchise Performance
u Examine the Bivariate Relationships. Use linear regression to build models
to answer these questions.
u Use the data to determine the relationship between SALES and FAMILIES (the
size of the customer base around each store).
u How is the size of the customer base around each store related to our expected
sales?
u Use the data to determine the relationship between SALES and SQFT (the
area each store occupies).
u Use the data to determine the relationship between SALES and INVENTORY
(the number of items in each store).
u Use the data to determine the relationship between SALES and ADVERTISING
(the amount of money each store spends on promotion).
u Use the data to determine the relationship between SALES and STORES (the
number of competing stores near by each store in our franchise).
Using SLR Analysis to Understand
Franchise Performance
u Perform the analysis and then give your answer to the questions
posted above, with you reason for the answers.
u The EXCEL data file has a data dictionary describing the meaning of
the variables.
Using Multivariate Regression
Models to Understand Franchise
Performance
u Use the same dataset in Franchises.xlxs.
u Build a model to answer the questions:
u Which collection of factors affect sales in our franchises and how?
u Can we create a model to design a store given certain demographics
so we can achieve a certain level of sales?
u Examine each of the variables and see if any variables need to be
transformed to improve the multi-variate model.
u Build a multivariate regression model that can be used to predict
store sales
Prediction Analysis
u Use the model to do some analysis.
u We just opened a store in a neighborhood with 5,000 families, the store is
5,000 sq ft, we are planning to spend $5,000 a month in advertising,
carry $250,000 in inventory and there are 5 competing stores in the
neighborhood.
u What are the projected sales?
Project Report Guidelines
u Include tables and visuals to support your analysis.
u Your project report should be 5-7 pages.
u You can use the following outline for your project report:
1. Introduction: introduce the problem, provide some background
2. Problem Statement: describe the problem you are proposing to
solve
u State the assumptions (if any)
u Data Sources: list the sources of data you expect to extract data for
developing the analysis.
3. Data Description: describe the data (explain any terms that are
specific to your dataset
Project Report Guidelines (cont.)
u Outline (cont.)
4. Methods: discuss your modelling approach
u What are the independent variables (co-variates)?
u What is the dependent variable?
u Provide as much detail about the various models you have in your
study.
u Did you need to transform any variables?
u How well do the variables fit the model?
u How well do the variables describe the model?
5. Results: what is the outcome of your analysis across the
different models?
6. Conclusion: Do you have a recommendation?
Project Submission Details
u Please submit only one report (in PDF format) for each group
u Your group report is due by Thursday April 14.
u Please submit your report to Blackboard -> Group Project 1 -> Submission
u Also submit your EXCEL file showing all the models you have developed
(one model per worksheet tab)
Powered by TCPDF (www.tcpdf.org) Powered by TCPDF (www.tcpdf.org)
