Using Hierarchical Clustering & PCA for Market Segmentation and Targeting
Javier Moreno for Group 6
11/01/2020
library(readr)
mydata <-read_csv('Coffee_Survey_Final_Clean.csv')## Rows: 28 Columns: 21## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## dbl (21): ID, Age, Occupation, Income, Zipcode, Buy_local_important, Drink_p...##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.###Building distance function and ploting the trees (dendrograms)
use = scale(mydata[,-c(1)], center = TRUE, scale = TRUE)
dist = dist(use)
d <- dist(as.matrix(dist)) # find distance matrix
seg.hclust <- hclust(d) # apply hirarchical clustering
library(ggplot2) # needs no introduction
plot(seg.hclust)
###Identifying clustering memberships for each cluster
groups.3 = cutree(seg.hclust,3)
table(groups.3) ## groups.3
## 1 2 3
## 8 18 2mydata$ID[groups.3 == 1]## [1] 1 3 13 14 18 20 22 28mydata$ID[groups.3 == 2]## [1] 2 4 5 6 8 9 10 11 12 15 16 19 21 23 24 25 26 27mydata$ID[groups.3 == 3]## [1] 7 17Identifying common features of each cluster using the aggregate function
#?aggregate
aggregate(mydata,list(groups.3),median)## Group.1 ID Age Occupation Income Zipcode Buy_local_important
## 1 1 16.0 3 2 5.5 10.5 2
## 2 2 13.5 3 2 5.0 8.0 2
## 3 3 12.0 4 3 6.0 7.0 1
## Drink_preference Cups_per_day Coffee_from Coffeeshop_visit Buy_reason_1
## 1 2 1.0 1.0 4 5.0
## 2 1 2.0 3.0 3 4.5
## 3 1 4.5 4.5 2 5.0
## Buy_reason_2 Buy_reason_3 Local_owned_important Organic_Important
## 1 7.0 5 0 2
## 2 5.0 5 1 2
## 3 5.5 6 0 0
## Selection_Important Social_cause_important Enviroment_important
## 1 0 0 0
## 2 1 1 1
## 3 0 1 0
## Drive thru_important Fast Service_important Non-coffee_purchases
## 1 0.0 0.0 3.0
## 2 1.0 1.0 3.0
## 3 1.5 0.5 2.5aggregate(mydata,list(groups.3),mean)## Group.1 ID Age Occupation Income Zipcode Buy_local_important
## 1 1 14.87500 3.125000 2.000000 4.875000 10.625000 2.375000
## 2 2 14.61111 2.944444 2.111111 4.944444 7.388889 2.277778
## 3 3 12.00000 4.000000 3.000000 6.000000 7.000000 1.000000
## Drink_preference Cups_per_day Coffee_from Coffeeshop_visit Buy_reason_1
## 1 2.125000 1.500000 2.0 3.625000 5.000000
## 2 1.277778 2.388889 2.5 3.055556 3.944444
## 3 1.000000 4.500000 4.5 2.000000 5.000000
## Buy_reason_2 Buy_reason_3 Local_owned_important Organic_Important
## 1 6.000000 4.875000 0.125000 1.500000
## 2 4.722222 5.333333 1.333333 1.611111
## 3 5.500000 6.000000 0.000000 0.000000
## Selection_Important Social_cause_important Enviroment_important
## 1 0.500000 0.000000 0.250000
## 2 1.444444 1.333333 1.333333
## 3 0.000000 1.000000 0.000000
## Drive thru_important Fast Service_important Non-coffee_purchases
## 1 0.375000 0.375000 2.750000
## 2 1.388889 1.222222 2.666667
## 3 1.500000 0.500000 2.500000aggregate(mydata[,-1],list(groups.3),median)## Group.1 Age Occupation Income Zipcode Buy_local_important Drink_preference
## 1 1 3 2 5.5 10.5 2 2
## 2 2 3 2 5.0 8.0 2 1
## 3 3 4 3 6.0 7.0 1 1
## Cups_per_day Coffee_from Coffeeshop_visit Buy_reason_1 Buy_reason_2
## 1 1.0 1.0 4 5.0 7.0
## 2 2.0 3.0 3 4.5 5.0
## 3 4.5 4.5 2 5.0 5.5
## Buy_reason_3 Local_owned_important Organic_Important Selection_Important
## 1 5 0 2 0
## 2 5 1 2 1
## 3 6 0 0 0
## Social_cause_important Enviroment_important Drive thru_important
## 1 0 0 0.0
## 2 1 1 1.0
## 3 1 0 1.5
## Fast Service_important Non-coffee_purchases
## 1 0.0 3.0
## 2 1.0 3.0
## 3 0.5 2.5aggregate(mydata[,-1],list(groups.3),mean)## Group.1 Age Occupation Income Zipcode Buy_local_important
## 1 1 3.125000 2.000000 4.875000 10.625000 2.375000
## 2 2 2.944444 2.111111 4.944444 7.388889 2.277778
## 3 3 4.000000 3.000000 6.000000 7.000000 1.000000
## Drink_preference Cups_per_day Coffee_from Coffeeshop_visit Buy_reason_1
## 1 2.125000 1.500000 2.0 3.625000 5.000000
## 2 1.277778 2.388889 2.5 3.055556 3.944444
## 3 1.000000 4.500000 4.5 2.000000 5.000000
## Buy_reason_2 Buy_reason_3 Local_owned_important Organic_Important
## 1 6.000000 4.875000 0.125000 1.500000
## 2 4.722222 5.333333 1.333333 1.611111
## 3 5.500000 6.000000 0.000000 0.000000
## Selection_Important Social_cause_important Enviroment_important
## 1 0.500000 0.000000 0.250000
## 2 1.444444 1.333333 1.333333
## 3 0.000000 1.000000 0.000000
## Drive thru_important Fast Service_important Non-coffee_purchases
## 1 0.375000 0.375000 2.750000
## 2 1.388889 1.222222 2.666667
## 3 1.500000 0.500000 2.500000cluster_means <- aggregate(mydata[,-1],list(groups.3),mean)Exporting cluster analysis results into excel from R Studio Cloud
write.csv(groups.3, "clusterID.csv")
write.csv(cluster_means, "cluster_means.csv")Principal Component Analysis (PCA)
Intro
Principal Component Analysis (PCA) involves the process of understanding different features in a dataset and can be used in conjunction with cluster analysis.
PCA is also a popular machine learning algorithm used for feature selection. Imagine if you have more than 100 features or factors. It is useful to select the most important features for further analysis.
The basic idea when using PCA as a tool for feature selection is to select variables according to the magnitude (from largest to smallest in absolute values) of their coefficients (loadings).
#install.packages('dplyr')
library(dplyr) ##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionlibrary(tidyr)
library(ggplot2)
library(ggfortify)
getwd() ## [1] "/cloud/project"library(readr)
mydata <-read_csv('Coffee_Survey_Final_Clean.csv')## Rows: 28 Columns: 21## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## dbl (21): ID, Age, Occupation, Income, Zipcode, Buy_local_important, Drink_p...##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.# read csv file #This allows you to read the data from my Github site.
fit <- kmeans(mydata[,-1], 3, iter.max=1000)
table(fit$cluster)##
## 1 2 3
## 9 9 10barplot(table(fit$cluster), col="#336699") 
pca <- prcomp(mydata[,-1], scale=TRUE)
pca_data <- mutate(fortify(pca), col=fit$cluster)
ggplot(pca_data) + geom_point(aes(x=PC1, y=PC2, fill=factor(col)),
size=3, col="#7f7f7f", shape=21) + theme_bw(base_family="Helvetica")
autoplot(fit, data=mydata[,-1], frame=TRUE, frame.type='norm')
names(pca)## [1] "sdev" "rotation" "center" "scale" "x"pca$center## Age Occupation Income
## 3.0714286 2.1428571 5.0000000
## Zipcode Buy_local_important Drink_preference
## 8.2857143 2.2142857 1.5000000
## Cups_per_day Coffee_from Coffeeshop_visit
## 2.2857143 2.5000000 3.1428571
## Buy_reason_1 Buy_reason_2 Buy_reason_3
## 4.3214286 5.1428571 5.2500000
## Local_owned_important Organic_Important Selection_Important
## 0.8928571 1.4642857 1.0714286
## Social_cause_important Enviroment_important Drive thru_important
## 0.9285714 0.9285714 1.1071429
## Fast Service_important Non-coffee_purchases
## 0.9285714 2.6785714pca$scale## Age Occupation Income
## 0.7663560 0.6506000 1.6996732
## Zipcode Buy_local_important Drink_preference
## 3.1839545 0.6862212 0.7934920
## Cups_per_day Coffee_from Coffeeshop_visit
## 1.4871584 1.6216133 1.6035675
## Buy_reason_1 Buy_reason_2 Buy_reason_3
## 2.1266139 2.3208282 2.4888641
## Local_owned_important Organic_Important Selection_Important
## 0.7859547 0.7444681 0.7663560
## Social_cause_important Enviroment_important Drive thru_important
## 0.7663560 0.7663560 0.6852567
## Fast Service_important Non-coffee_purchases
## 0.6042180 0.5479640pca$rotation## PC1 PC2 PC3 PC4
## Age 0.037705866 0.36200898 -0.25492300 0.277720594
## Occupation 0.200337337 0.31411542 -0.30984707 0.018995509
## Income 0.088140678 0.33361268 0.12170734 0.022700383
## Zipcode -0.245378290 -0.07390904 0.01157851 0.083895243
## Buy_local_important -0.026460888 -0.11675408 0.34865242 -0.170547295
## Drink_preference -0.310902984 -0.31227903 -0.18312270 0.124427011
## Cups_per_day 0.193843848 0.30722977 0.08276182 0.046721259
## Coffee_from 0.039454241 -0.01453535 -0.24394395 0.525338263
## Coffeeshop_visit -0.134835477 -0.25741700 -0.13332409 0.332938205
## Buy_reason_1 -0.246046352 -0.07242288 0.33037781 0.288070391
## Buy_reason_2 -0.160141513 0.27896225 0.10750443 -0.137031439
## Buy_reason_3 0.043037046 -0.08494933 -0.32998342 -0.399981137
## Local_owned_important 0.380602838 -0.23108077 -0.11396336 0.018568791
## Organic_Important 0.042944242 -0.31134255 -0.19266196 -0.092094952
## Selection_Important 0.244890127 -0.28224032 0.29697831 -0.089602168
## Social_cause_important 0.402790882 -0.11629244 -0.16800844 0.002491132
## Enviroment_important 0.361065745 -0.20281076 -0.04124596 0.058871040
## Drive thru_important 0.233340441 -0.01187485 0.27826266 0.339608416
## Fast Service_important 0.306523635 0.01683559 0.32822086 0.127844307
## Non-coffee_purchases -0.008169555 -0.07843671 -0.04302496 0.260048637
## PC5 PC6 PC7 PC8
## Age -0.10026479 0.04587032 -0.354413036 0.013497404
## Occupation -0.28449398 0.20092746 0.015387211 -0.042227295
## Income 0.43364359 0.19707819 0.093696135 -0.052223560
## Zipcode 0.49236492 0.06709525 -0.159193343 0.427339154
## Buy_local_important -0.34920793 0.41135090 -0.239133857 0.060734360
## Drink_preference -0.03869161 0.00841105 0.149242416 0.139436264
## Cups_per_day 0.24361892 -0.04470592 0.112006960 0.289751915
## Coffee_from 0.02137129 -0.11940973 0.240584200 0.168203235
## Coffeeshop_visit 0.08385195 0.15564659 -0.429920041 -0.091561674
## Buy_reason_1 -0.05919088 -0.18498475 0.183625380 -0.332285018
## Buy_reason_2 -0.25172810 -0.16534818 0.172780287 0.451467741
## Buy_reason_3 0.04515812 -0.04553451 0.233312957 -0.006057242
## Local_owned_important -0.12920666 -0.05644509 0.028503482 0.017101234
## Organic_Important -0.15131322 -0.28427682 -0.118538135 0.416527552
## Selection_Important 0.23178344 0.07678892 0.198231648 0.056868447
## Social_cause_important 0.13955443 -0.18289523 0.009112269 -0.227950246
## Enviroment_important 0.14248083 0.34103969 -0.067356021 0.113250383
## Drive thru_important -0.14658824 -0.30313176 0.125091099 -0.002537500
## Fast Service_important -0.15550122 -0.07964847 -0.268048283 0.301539664
## Non-coffee_purchases -0.20147072 0.55036249 0.490952126 0.148023845
## PC9 PC10 PC11 PC12
## Age -0.00818993 0.245370167 0.185110213 -0.156171178
## Occupation 0.02219942 -0.161741078 -0.103271884 -0.310455438
## Income -0.15440313 0.563342524 -0.007512520 -0.009963682
## Zipcode 0.28865191 -0.166003308 -0.317856229 -0.160078815
## Buy_local_important 0.22464386 0.178055013 0.242307336 -0.122122585
## Drink_preference 0.12931668 0.076219044 0.046339274 0.003348516
## Cups_per_day -0.17030342 -0.502381508 0.495013123 0.103776743
## Coffee_from 0.23370925 0.230238937 0.051923083 -0.106748629
## Coffeeshop_visit -0.19466633 -0.004078856 0.276863994 0.494050063
## Buy_reason_1 -0.17527861 0.026570780 0.024293198 -0.072474243
## Buy_reason_2 -0.18110563 0.312331770 -0.175764318 0.458353697
## Buy_reason_3 0.44971160 0.147290282 0.387028739 0.237672332
## Local_owned_important -0.09699525 -0.047011417 -0.390832062 0.126878622
## Organic_Important -0.44289684 0.169577393 0.216738239 -0.369202691
## Selection_Important -0.14757391 0.138250884 0.196331791 -0.247156220
## Social_cause_important -0.07938053 0.054716390 -0.063834313 0.195385685
## Enviroment_important 0.03163209 0.158222377 -0.097270273 0.068954133
## Drive thru_important 0.37238777 0.057381043 0.164406408 -0.021977914
## Fast Service_important 0.18538928 -0.058236195 -0.092795013 0.176425196
## Non-coffee_purchases -0.17356306 -0.145089658 0.003117891 0.111647241
## PC13 PC14 PC15 PC16
## Age 0.05893407 -0.350876459 0.2383626129 -0.142282435
## Occupation -0.28234323 0.198198204 0.1256690699 -0.188306696
## Income 0.19625774 0.335203520 0.2382062254 0.196182435
## Zipcode 0.08861019 -0.030922765 0.1922491828 -0.338387362
## Buy_local_important -0.07879331 0.365228920 -0.0001364952 -0.172562135
## Drink_preference -0.37472890 0.237244225 0.4653164448 0.280230822
## Cups_per_day -0.15997922 0.253950399 0.0832004198 -0.065943474
## Coffee_from -0.04075050 0.288493823 -0.5794524206 -0.022198725
## Coffeeshop_visit -0.06879855 0.051134472 -0.0197038575 0.005143447
## Buy_reason_1 0.07300437 0.148876031 0.1255075769 -0.628573490
## Buy_reason_2 -0.30672729 -0.104798964 -0.0155278631 -0.204808722
## Buy_reason_3 0.26985533 -0.006679436 0.0766677445 -0.284036389
## Local_owned_important 0.06664852 0.307540954 0.2368766529 -0.004930643
## Organic_Important 0.25688019 0.035019556 0.0749720675 -0.059363400
## Selection_Important -0.28051652 -0.232249219 -0.0985888560 0.007283355
## Social_cause_important -0.04754710 0.085869406 0.1439263535 -0.249445902
## Enviroment_important -0.30674223 -0.250082567 -0.0830926695 -0.252314334
## Drive thru_important -0.06936985 -0.264451005 0.3685396705 0.161688910
## Fast Service_important 0.33407748 0.131324982 -0.0533857033 0.068523021
## Non-coffee_purchases 0.39711369 -0.210332846 0.1203348785 0.008612881
## PC17 PC18 PC19 PC20
## Age 0.274356487 -0.13606971 0.300198788 -0.28457387
## Occupation -0.079263490 0.48249132 -0.307678403 0.06166248
## Income -0.149522281 0.08810182 -0.076150659 0.06193634
## Zipcode 0.135157817 0.10550146 0.014001949 0.21181452
## Buy_local_important 0.182199182 -0.21401806 0.119188840 0.25419338
## Drink_preference 0.071545005 -0.14938324 -0.162517812 -0.37845678
## Cups_per_day -0.057011778 -0.18459811 0.153089685 -0.04283930
## Coffee_from 0.089193548 -0.04172606 0.086517081 0.04247881
## Coffeeshop_visit -0.036558721 0.39909903 -0.026442561 0.18636450
## Buy_reason_1 -0.163248010 0.03763563 -0.003691161 -0.23775154
## Buy_reason_2 0.089682274 0.08026788 0.008909056 0.09310075
## Buy_reason_3 -0.089533424 0.21194787 0.057261614 -0.13274297
## Local_owned_important -0.009671652 0.11642900 0.645886392 -0.03134568
## Organic_Important -0.184747893 -0.02693577 -0.140793209 0.15851543
## Selection_Important 0.385932343 0.42869891 0.102838111 -0.18912081
## Social_cause_important 0.498209014 -0.31961928 -0.391558958 0.20791125
## Enviroment_important -0.539155725 -0.28926403 -0.075967570 -0.14757108
## Drive thru_important -0.194092336 0.08584015 0.001137816 0.41519315
## Fast Service_important 0.056044403 0.14132392 -0.346667512 -0.48083004
## Non-coffee_purchases 0.146190235 -0.05061233 -0.082293095 0.08379117dim(pca$x)## [1] 28 20biplot(pca, scale=0)
pca$rotation=-pca$rotation
pca$x=-pca$x
biplot(pca, scale=0)
pca$sdev## [1] 2.0529879 1.7193528 1.5565201 1.4222525 1.1574714 1.1394458 1.0415987
## [8] 0.9746165 0.8477479 0.7977981 0.7702227 0.7035045 0.5621017 0.5431589
## [15] 0.4719661 0.4049420 0.3188104 0.2941911 0.2296571 0.1693139pca.var=pca$sdev^2
pca.var## [1] 4.21475925 2.95617418 2.42275494 2.02280223 1.33973996 1.29833662
## [7] 1.08492777 0.94987733 0.71867654 0.63648174 0.59324295 0.49491851
## [13] 0.31595829 0.29502164 0.22275196 0.16397804 0.10164008 0.08654839
## [19] 0.05274241 0.02866719pve=pca.var/sum(pca.var)
pve## [1] 0.210737962 0.147808709 0.121137747 0.101140112 0.066986998 0.064916831
## [7] 0.054246388 0.047493866 0.035933827 0.031824087 0.029662148 0.024745926
## [13] 0.015797914 0.014751082 0.011137598 0.008198902 0.005082004 0.004327419
## [19] 0.002637120 0.001433359plot(pve, xlab="Principal Component", ylab="Proportion of Variance Explained", ylim=c(0,1),type='b')
plot(cumsum(pve), xlab="Principal Component", ylab="Cumulative Proportion of Variance Explained", ylim=c(0,1),type='b')
write.csv(pca_data, "pca_data.csv")References
Cluster analysis - reading (p.385-p.399) https://www.statlearning.com/
Hint:you can download the free version of this book from this website.
Comparison of similarity coefficients used for cluster analysis with dominant markers in maize (Zea mays L) https://www.scielo.br/scielo.php?script=sci_arttext&pid=S1415-47572004000100014&lng=en&nrm=iso
Principal Component Methods in R: Practical Guide http://www.sthda.com/english/articles/31-principal-component-methods-in-r-practical-guide/118-principal-component-analysis-in-r-prcomp-vs-princomp/
Principal component analysis - reading (p.404-p.405) https://www.statlearning.com/
Hint:you can download the free version from this website.
Principal Component Methods in R: Practical Guide http://www.sthda.com/english/articles/31-principal-component-methods-in-r-practical-guide/118-principal-component-analysis-in-r-prcomp-vs-princomp/