Identifying Candidate Sites for T&T Supermarket's New Store

The assignment is related to Cluster Analysis, specifically identifying candidate sites for a new store of T&T Supermarket in the Toronto region. The task involves performing a K-means cluster analysis on census data using variables such as Chinese-vm, established immigrants, owner, average household income, percentage of households not low-income, and education levels. The goal is to identify areas with high concentration of affluent ethnic Chinese consumers. The assignment requires writing an overview of Cluster Analysis, performing the K-means analysis, mapping the clusters, adding existing T&T stores, suggesting 1-3 candidate sites for the new store, and describing and comparing the candidate sites.

GEO 561: Multivariate Analytical Techniques (F2016)
Instructor: Shuguang Wang
Assignment 2
Factor Analysis and Cluster Analysis (50 Marks)
Due date: Dec 2, 2016
Factor Analysis (25 marks):
1. Write an overview of Principal Component Analysis and Factor Analysis (4-5 pages)
with illustrations and references
2. The provided data set (“Toronto HH Income by neighborhood”) contains 10 variables for
333 neighborhoods in the Toronto CMA from the 2011 census.
 Use 8 of them (highlighted in yellow in the table below) to perform a Factor Analysis.
Name the extracted factors with appropriate descriptive labels, and interpret the output in
writing.
 Using the extracted (and saved) factor scores to perform a regression analysis, with
“AVHHIN11” as DEPENDENT variable, and the factor scores as INDEPENDENT
variables; use the regression model to estimate AVHH11 for the neighborhood that is
assigned to you.
 Construct a regression model using “AVHHIN11” as DEPENDENT variable, and the
ORIGINAL variables as INDEPENDENTS; use the regression model to estimate
AVHH11 for the neighborhood that is assigned to you.
 compare the two estimated average household incomes.
List of variables in “Toronto HH Income by neighborhood”.
Field Description Expected effects
TOTPOP total population (2011)
PUNIV % population with university or higher degree +
PIAF01 % immigrants after 2000 -
PMT % population with mother tongue not official languages -
PVM % visible minority -
PWHITE % population being white collar workers +
PBLUE % population being blue collar workers -
AVHHIN11 average household income (2011)
PMARRIED % married population +
PLPF % lone parent family -

3. The second data set (“Toronto house value by neighborhood”) contains 13 variables for 333
neighborhoods in the Toronto CMA from the 2011 census.
 Use 9 of them (highlighted in yellow in the table below) to perform a Factor Analysis.
Name the extracted factors with appropriate descriptive labels, and interpret the output in
writing.
 Use the extracted (and saved) factor scores to perform a regression analysis, with
“AVDVAL11” as DEPENDENT variable, and the factor scores as INDEPENDENT
variables; use the regression model to estimate AVDVAL11 for the neighborhood that is
assigned to you
 Construct a regression model using AVDVAL11 as DEPENDENT variable, and the
ORIGINAL variables as INDEPENDENTS; use the regression model to estimate
AVDVAL11 for the neighborhood that is assigned to you.
 compare the two estimated average household incomes.
List of variables in the second date set
Field Description
Expected
effects
TOTPOP total population (2011)
TOTDWL Total number of dwellings
POWNED % owned house
PHBF80 % house built before 1980
PSINGDET % single-detached +
AVRM average number of rooms -
MJ house with major repair +
AVDVAL11 average house value (2011)
POP_GRW population growth rate +
R_SR_C number of regional and super regional shopping centres +
GROCER number of supermarket stores +
GREENCAP green space per capita _

Neighborhood assigned to you.
Neighborhood Student
Oakville - 2 Bautista,Alexandra A
Mississauga - 7 Bocknek,Daniel
Lawrence Park South Borden,Andrew
Rosedale-Moore Park Codinera,Nikki
Kingsway South Connor,Cody
Forest Hill South Da Silva,Melissa
Markham - 20 Defrias,Nicolas
Mississauga - 8 Dhawan,Megha
Mississauga - 20 Garcha,Bharvat
Vaughan - 6 Grandison,Andrew
St.Andrew-Windfields
Juarez,Anna-Christina
Juarez
Lawrence Park North Khan-Yusafzai,Asfandyar
Mississauga - 5 Matadeen,Renacin Jordan
Vaughan - 3 Omar,Mahomed Mahir
Bedford Park-Nortown Shahid,Mirza Ammar
Casa Loma Shatrov,Alexander
Princess-Rosethorn Soto,Mathew
Leaside-Bennington Stefanova,Desislava
Oakville - 11 Tahir,Zara
Richmond Hill - 6 Bagchi,Priyanka

Cluster Analysis (25 marks)
T&T Supermarket (headquartered in Vancouver) is a food retailer targeting ethnic Chinese
consumers. Unlike the other Chinese food retailers, T&T focuses on the affluent Chinese with
high household income. After being acquired by Loblaw in 2009, it has expanded to operate 7
stores in the Toronto region. Recently, it “plans” to open another store in the region. The most
important location criterion is a spatial concentration of affluent ethnic Chinese in the
surrounding trade area of the potential sites. You are invited to assist T&T to identify 1-3
candidate sites for the new store.
Tasks:
1. Write an overview of Cluster Analysis with references used (2-3 pages).
2. Perform a K-means Cluster Analysis on the provided census data to identify areas of
concentration of affluent ethnic Chinese in the Toronto region.
Use the following census variables as classification variables (note: two of them need to
be calculated by yourself from the original variables in the data set – known as “derived
variables”)
 Chinese-vm
 Established immigrants (im – im_06_11)
 Owner
 Avg_hh_inc
 Percentage of household not low income (1 – pop_prv)
 Bach_25
 Uni_o_25
Experiment with K = 3, K = 4 and k=5 to find the most suitable classification
3. Join the SPSS data table (with cluster ID) with the MapInfo table, and map the clusters.
4. Add the existing T&T stores to the map
5. Suggest 1-3 locations for the new store.
6. Describe and compare the candidate sites in writing (1-2 pages)

Toronto CMA 2011 Census Variable Definitions
Database
Field Original Field
ctuid Census tract ID
totpop_cz Total population in private households by citizenship
can_cz Canadian citizens
im Immigrants
im06_11 Recent immigrants who arrived between 2006 and 2011
totvm Total visible minority population
s_asian_vm South Asian
chinese_vm Chinese
black_vm Black
filip_vm Filipino
latin_vm Latin American
arab_vm Arab
se_asia_vm Southeast Asian
w_asia_vm West Asian
korean_vm Korean
japan_vm Japanese
unemploye
d Number of unemployed persons
unemploy Unemployment rate
household Total number of private households by tenure
owner Home owner
renter Home renter
Avg_val Average value of dwellings
subsid % of tenant households in subsidized housing
med_inc Median personal income $
avg_inc Average personal income $
hh_inc Household income in 2010 of private households
med_hh_in
c Median household total income $
avg_hh_inc Average household total income $
pop_prv Prevalence of low income in 2010 based on after-tax low-income measure %
pop_dgr_25
Total population aged 25 to 64 years by highest certificate, diploma or
degree
bach_25 population aged 25 to 64 years with Bachelor's degree
uni_o_25
population aged 25 to 64 years with University certificate, diploma or
degree above bachelor level