-
Notifications
You must be signed in to change notification settings - Fork 0
/
script.R
290 lines (215 loc) · 8.84 KB
/
script.R
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
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
# Introduction to the "Google Analytics Customer Revenue Prediction".
library(tidyverse)
library(tictoc)
library(jsonlite)
library(magrittr)
library(dplyr)
library(naniar)
library(reshape2)
library(ggplot2)
library(Matrix)
library(xgboost)
# Function to visualize the dataset upload
load_info <- function(file){
message(paste("Starting the upload..."))
tic()
train <- read.csv(file, header = TRUE)
message(paste("------------------"))
message(paste("End of the upload..."))
toc()
message(paste("------------------"))
return(train)
}
# Function to remove NAs
remove_nas <- function(y, list){
is_na_val <- function(x) x %in% list
y <- mutate_if(y, is.factor,is.character)
y <- mutate_if(y, is.logical,is.character)
y <- y %>% mutate_all(funs(ifelse(is_na_val(.), NA, .)))
y[is.na(y)] <- 0
return(y)
}
# Function to clean the JSON format
clean_json <- function(data){
message(paste("Starting the transformation for JSON format..."))
library(jsonlite)
flatten_json <- . %>%
str_c(., collapse = ",") %>%
str_c("[", ., "]") %>%
fromJSON(flatten = T)
parse <- . %>%
bind_cols(flatten_json(.$device)) %>%
bind_cols(flatten_json(.$geoNetwork)) %>%
bind_cols(flatten_json(.$trafficSource)) %>%
bind_cols(flatten_json(.$totals)) %>%
select(-device, -geoNetwork, -trafficSource, -totals)
tic()
train <- parse(data)
message(paste("------------------"))
message(paste("End of the transformation."))
toc()
message(paste("------------------"))
return(train)
}
# Function to remove na in variables some specific variables
remove_subconts <- function(y, list){
is_na_val <- function(x) x %in% list
y <- mutate_if(y, is.factor,is.character)
y <- mutate_if(y, is.logical,is.character)
y <- y %>% mutate_all(funs(ifelse(!is_na_val(.), NA, .)))
y[is.na(y)] <- "other"
return(y)
}
# Function to clean
trash_info <- function(x) {
trash_list <- c()
for (header in colnames(x)) {
if (nrow(unique(x[paste(as.character(header))])) == 1) {
trash_list <- c(trash_list,as.character(header))
}
}
return(trash_list)
}
list_not_set <- c("not available in demo dataset", "(not provided)",
"(not set)", "<NA>", "unknown.unknown","(none)")
# Once defined the functions, we are going to upload the datasets and make use of the functions above to clean de JSON format.
# Loading dataset and cleaninng
train <- load_info(".../train.csv")
test <- load_info(".../test.csv")
# Cleaning the JSON format
train <- clean_json(train)
test <- clean_json(test)
# Once we have the dataset uploaded and ready to work on it, it's time to visualize the variables.
png("g1.png")
g1 <- head(train,200000) %>%
is.na %>% melt %>%
ggplot(data = .,aes(y = Var1,x = Var2)) +
geom_raster(aes(fill = value)) + coord_flip() +
scale_fill_grey(name = "",labels = c("Present","Missing")) +
labs(x = "Observation",y = "Variables")
print(g1)
dev.off()
knitr::include_graphics("g1.png")
# Cleaning trash
train.tash <- trash_info(train)
test.tash <- trash_info(test)
trash.list <- names(train) %in% trash_info(train)
trash.list.test <- names(test) %in% trash_info(test)
train <- train[!trash.list]
test <- test[!trash.list.test]
train <- remove_nas(train,list_not_set)
test <- remove_nas(test,list_not_set)
# Defining the dataset as a tibble
train.tib <- as.tibble(train)
test.tib <- as.tibble(test)
print(train.tib, n=7, width=Inf)
# Creating master table
MT.train <- select(train.tib,
#############################################################################
## add here new variables, they are already clean (NA -> 0) ##
#############################################################################
visitNumber,
medium,
isTrueDirect,
hits,
pageviews,
bounces,
newVisits,
country,
operatingSystem,
deviceCategory,
browser,
subContinent,
date,
transactionRevenue)
MT.test <- select(test.tib,
#############################################################################
## add here new variables, they are already clean (NA -> 0) ##
#############################################################################
visitNumber,
medium,
isTrueDirect,
hits,
pageviews,
bounces,
newVisits,
country,
subContinent,
operatingSystem,
deviceCategory,
browser,
date)
###########################################################################
############################################################################
## Aqui estic seleccionant els subcontinents mes rellevants ##
###########################################################################
MT.train$aux_ind <- 0
MT.test$aux_ind <- 0
MT.train$aux_ind[MT.train$subContinent == "Northern America"] <- 1
MT.train$aux_ind[MT.train$subContinent == "South America"] <- 1
MT.train$aux_ind[MT.train$subContinent == "Eastern Asia"] <- 1
MT.test$aux_ind[MT.test$subContinent == "Northern America"] <- 1
MT.test$aux_ind[MT.test$subContinent == "South America"] <- 1
MT.test$aux_ind[MT.test$subContinent == "Eastern Asia"] <- 1
MT.train$subContinent[MT.train$aux_ind == 0] <- "other"
MT.test$subContinent[MT.test$aux_ind == 0] <- "other"
MT.train <- select(MT.train, -aux_ind)
MT.test <- select(MT.test, -aux_ind)
#############################################################################
MT.train <- mutate(MT.train,
#############################################################################
## convert new varibles into your format ##
#############################################################################
transactionRevenue = as.double(transactionRevenue),
visitNumber = as.integer(visitNumber),
medium = as.factor(medium),
isTrueDirect = as.integer(isTrueDirect),
hits = as.integer(hits),
pageviews = as.integer(pageviews),
bounces = as.integer(bounces),
newVisits = as.integer(newVisits),
country = as.factor(country),
subContinent = as.factor(subContinent),
operatingSystem = as.factor(operatingSystem),
deviceCategory=as.factor(deviceCategory),
browser=as.factor(browser),
date = as.Date(as.character(date),"%Y%m%d"))
MT.test <- mutate(MT.test,
#############################################################################
## convert new varibles into your format ##
#############################################################################
visitNumber = as.integer(visitNumber),
medium = as.factor(medium),
isTrueDirect = as.integer(isTrueDirect),
hits = as.integer(hits),
pageviews = as.integer(pageviews),
bounces = as.integer(bounces),
newVisits = as.integer(newVisits),
country = as.factor(country),
subContinent = as.factor(subContinent),
operatingSystem = as.factor(operatingSystem),
deviceCategory=as.factor(deviceCategory),
browser=as.factor(browser),
date = as.Date(as.character(date),"%Y%m%d"))
#############################################################################
## Conversion of transaction revenue ##
#############################################################################
MT.train$transaction[MT.train$transactionRevenue > 0] <- 1
MT.train$transaction[MT.train$transactionRevenue == 0] <- 0
MT.train <- select(MT.train , -transactionRevenue)
# Logistic Regression
# Subsamples
set.seed(100)
ind<-sample(2, nrow(MT.train), replace = T, prob=c(0.7, 0.3))
train1<-MT.train[ind==1,]
test2<-MT.train[ind==2,]
#Logistic function
logis<-glm(transaction ~ medium + deviceCategory + hits + pageviews + subContinent + newVisits + isTrueDirect + visitNumber, data=train1, family="binomial")
summary(logis)
#Predictions
p1<-predict(logis, train1, type = "response")
head(p1)
#Classification error:
pred1<-ifelse(p1>0.1, 1, 0)
tab1<-table(Predicted= pred1, Actual=train1$transaction)
tab1