Digital-Marketing-/Code_practice3.R at master · 100rabh1401/Digital-Marketing- · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
remove(list=ls())

#Set working directory
setwd("F:/Cosmodea")

#Attach Libraries
library(data.table)
library(tm)
library(FSelector)
library(stringr)
library(dplyr)
library(caret)
library(lubridate)
library(SiteCatalyst)
library(RTextTools)
library(cluster)

#Import file
Report <- fread("Search term report(1).csv")
Report <- Report[-17743,]

#Summary of file
summary(Report)

#Change the column names
setnames(Report,
         c("Match type", "Search term", "Added/Excluded", "Campaign",
           "Ad group", "Clicks", "Impressions", "CTR", "Avg. CPC", "Cost",
           "Avg. position", "Conversions", "Cost / conv.", "Conv. rate",
           "All conv.", "View-through conv."),
         c("Match_type", "Search_term", "Added/Excluded", "Campaign",
           "Ad_group", "Clicks", "Impressions", "CTR", "Avg.CPC", "Cost",
           "Avg.position", "Conversions", "Cost/conv", "Conv.rate",
           "All_conv", "View_through_conv"))

#Convert CTR, Cost, Conv.rate & Conv.rate into numeric
Report <- Report %>%
  mutate(
    CTR = as.numeric(sub("%","",CTR)),
    Cost = as.numeric(str_replace_all(Cost,",","")),
    `Cost/conv` = as.numeric(str_replace_all(`Cost/conv`,",","")),
    Conv.rate = as.numeric(sub("%","",Conv.rate))
   )

#Convert character variables into factor
Report[,c(1:5)] <- lapply(Report[,c(1:5)], factor)

unique_search <- unique(Report$Search_term)
##write.csv(unique_search,"Search_term.csv")
##Removing Factor columns
Report_new <- Report[,-c(1,3,4,5)]

#sample.df.cor <- cor(Report[,-c(1:5)])
#sample.df.highly.cor <- findCorrelation(sample.df.cor, cutoff=.7)

#Check Structure of file
str(Report)

#Take one row out from the data table
df <- as.character(Report$Search_term)

#COnvert data table into vector
df_matrix <- create_matrix(df,
                     stemWords=TRUE,
                     minWordLength=1,
                     removePunctuation= TRUE,
                     weighting=weightTf)
names(df_matrix)

unique_words <- findFreqTerms(df_matrix)
#binding <- cbind(df_matrix$i,df_matrix$j,df_matrix$v)

#Build model
kmeans.model_7 <- kmeans(df_matrix, 7)

#Build plot for the cluster
clusplot(as.data.frame(df_matrix), kmeans.model_7$cluster,
         main='Representation of the Cluster solution',
         color=TRUE, shade=TRUE,
         labels=2, lines=0)

#Making the cluster
kw_with_cluster <- as.data.frame(cbind(df_vec, kmeans.model_7$cluster))
names(kw_with_cluster) <- c("keyword", "kmeans_20")

cluster1 <- subset(kw_with_cluster, subset=kmeans_20 == 1)
cluster2 <- subset(kw_with_cluster, subset=kmeans_20 == 2)
cluster3 <- subset(kw_with_cluster, subset=kmeans_20 == 3)
cluster4 <- subset(kw_with_cluster, subset=kmeans_20 == 4)
cluster5 <- subset(kw_with_cluster, subset=kmeans_20 == 5)
cluster6 <- subset(kw_with_cluster, subset=kmeans_20 == 6)
cluster7 <- subset(kw_with_cluster, subset=kmeans_20 == 7)
cluster8 <- subset(kw_with_cluster, subset=kmeans_20 == 8)
cluster9 <- subset(kw_with_cluster, subset=kmeans_20 == 9)
cluster10 <- subset(kw_with_cluster, subset=kmeans_20 == 10)
cluster11 <- subset(kw_with_cluster, subset=kmeans_20 == 11)
cluster12 <- subset(kw_with_cluster, subset=kmeans_20 == 12)
cluster13 <- subset(kw_with_cluster, subset=kmeans_20 == 13)
cluster14 <- subset(kw_with_cluster, subset=kmeans_20 == 14)
cluster15 <- subset(kw_with_cluster, subset=kmeans_20 == 15)
cluster16 <- subset(kw_with_cluster, subset=kmeans_20 == 16)
cluster17 <- subset(kw_with_cluster, subset=kmeans_20 == 17)
cluster18 <- subset(kw_with_cluster, subset=kmeans_20 == 18)
cluster19 <- subset(kw_with_cluster, subset=kmeans_20 == 19)
cluster20 <- subset(kw_with_cluster, subset=kmeans_20 == 20)

bb <- rbind(cluster1,cluster2,cluster3,cluster4,cluster5,cluster6,cluster7,cluster8,
            cluster9,cluster10,cluster11,cluster12,cluster13,cluster14,cluster15,
            cluster16,cluster17,cluster18,cluster19,cluster20)

#Export the clusters file
write.csv(bb,"alldata_in_clusters_20.csv")


#Check for the appropriate value of k by plotting
wtss<-c()
for(i in 1:20)
{
  model1<-kmeans(df_matrix,i)
  wtss<-c(wtss,model1$tot.withinss)
}
plot(wtss, type="b")

#Develop the model where distance calculator is Hamming
kmeans_133 <- kmeans(df_matrix, 133)

#Making the cluster
kw_with_cluster <- as.data.frame(cbind(df_vec, kmeans_133$cluster))
names(kw_with_cluster) <- c("keyword", "kmeans_20")