title: “3_22 HW” author: “Kyle” date: “2019年3月22日” output: html_document — —

#讀取資料
library(readr)
data7 <- read_csv("201807_data.csv")

## Parsed with column specification:
## cols(
##   Date = col_character(),
##   Page_Name = col_character(),
##   Page_ID = col_double(),
##   Link = col_logical(),
##   Type = col_character(),
##   All_Reaction_Count = col_double(),
##   LIKE_COUNT = col_double(),
##   WOW_COUNT = col_double(),
##   LOVE_COUNT = col_double(),
##   HAHA_COUNT = col_double(),
##   SAD_COUNT = col_double(),
##   ANGRY_COUNT = col_double(),
##   Comment_Count = col_double(),
##   Share_Count = col_double(),
##   Message = col_character(),
##   Link_Title = col_character(),
##   `Link Description` = col_character(),
##   created_time = col_double()
## )

data8 <- read_csv("201808_data.csv")

## Parsed with column specification:
## cols(
##   Date = col_character(),
##   Page_Name = col_character(),
##   Page_ID = col_double(),
##   Link = col_logical(),
##   Type = col_character(),
##   All_Reaction_Count = col_double(),
##   LIKE_COUNT = col_double(),
##   WOW_COUNT = col_double(),
##   LOVE_COUNT = col_double(),
##   HAHA_COUNT = col_double(),
##   SAD_COUNT = col_double(),
##   ANGRY_COUNT = col_double(),
##   Comment_Count = col_double(),
##   Share_Count = col_double(),
##   Message = col_character(),
##   Link_Title = col_character(),
##   `Link Description` = col_character(),
##   created_time = col_double()
## )

data9 <- read_csv("201809_data.csv")

## Parsed with column specification:
## cols(
##   Date = col_character(),
##   Page_Name = col_character(),
##   Page_ID = col_double(),
##   Link = col_logical(),
##   Type = col_character(),
##   All_Reaction_Count = col_double(),
##   LIKE_COUNT = col_double(),
##   WOW_COUNT = col_double(),
##   LOVE_COUNT = col_double(),
##   HAHA_COUNT = col_double(),
##   SAD_COUNT = col_double(),
##   ANGRY_COUNT = col_double(),
##   Comment_Count = col_double(),
##   Share_Count = col_double(),
##   Message = col_character(),
##   Link_Title = col_character(),
##   `Link Description` = col_character(),
##   created_time = col_double()
## )

data10 <- read_csv("201810_data.csv")

## Parsed with column specification:
## cols(
##   Date = col_character(),
##   Page_Name = col_character(),
##   Page_ID = col_double(),
##   Link = col_logical(),
##   Type = col_character(),
##   All_Reaction_Count = col_double(),
##   LIKE_COUNT = col_double(),
##   WOW_COUNT = col_double(),
##   LOVE_COUNT = col_double(),
##   HAHA_COUNT = col_double(),
##   SAD_COUNT = col_double(),
##   ANGRY_COUNT = col_double(),
##   Comment_Count = col_double(),
##   Share_Count = col_double(),
##   Message = col_character(),
##   Link_Title = col_character(),
##   `Link Description` = col_character(),
##   created_time = col_double()
## )

data11 = read_csv("201811_data.csv")

## Parsed with column specification:
## cols(
##   Date = col_character(),
##   Page_Name = col_character(),
##   Page_ID = col_double(),
##   Link = col_logical(),
##   Type = col_character(),
##   All_Reaction_Count = col_double(),
##   LIKE_COUNT = col_double(),
##   WOW_COUNT = col_double(),
##   LOVE_COUNT = col_double(),
##   HAHA_COUNT = col_double(),
##   SAD_COUNT = col_double(),
##   ANGRY_COUNT = col_double(),
##   Comment_Count = col_double(),
##   Share_Count = col_double(),
##   Message = col_character(),
##   Link_Title = col_character(),
##   `Link Description` = col_character(),
##   created_time = col_double()
## )

data=rbind(data7, data8,data9,data10,data11)
rm(data7,data8,data9,data10,data11)

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, union

Su = filter(data, grepl("蘇貞昌", data$Page_Name) == T&grepl("蘇貞昌後援會", data$Page_Name) == F) 
Hou = filter(data, grepl("侯友宜", data$Page_Name) == T&grepl("鐵漢柔情 力挺侯友宜", data$Page_Name) == F) 

#設定時間
Su$Date = as.POSIXct(Su$Date,format = "%Y/%m/%d %H:%M:%S")
Hou$Date = as.POSIXct(Hou$Date,format = "%Y/%m/%d %H:%M:%S")

#把11/24之後的貼文刪除以及加一個只有日期沒有時間的欄位
library(ggplot2)
Pei = rbind(Hou,Su)

Pei = filter(Pei,Pei$Date<"2018/11/24")
Pei$cDate = as.character(Pei$Date)
Pei$cDate = sapply(Pei$cDate,function(x) strsplit(x,"[ ]") %>% unlist %>% .[1])
Pei$cDate = Pei$cDate %>% as.Date("%Y-%m-%d")

#兩候選人總Po文類型比較
Peitype=Pei%>%group_by(Type)%>%count()
label_value <- paste('(', round(Peitype$n/sum(Peitype$n) * 100, 1), '%)', sep = '')
label=paste(Peitype$Type,label_value,sep = "")
ggplot(Peitype, aes(x="", y=n, fill=Type)) +
       geom_bar(width = 1, stat = "identity") +
       coord_polar("y", start=0)+
       labs(x = '', y = '', title = '')+
       theme(axis.text = element_blank())+
       scale_fill_discrete(labels = label)

#蘇貞昌貼文類型分析
Sutype=Su%>%group_by(Type)%>%count()
label_value <- paste('(', round(Sutype$n/sum(Sutype$n) * 100, 1), '%)', sep = '')
label=paste(Sutype$Type,label_value,sep = "")
ggplot(Sutype, aes(x="", y=n, fill=Type)) +
       geom_bar(width = 1, stat = "identity") +
       coord_polar("y", start=0)+
       labs(x = '', y = '', title = '')+
       theme(axis.text = element_blank())+
       scale_fill_discrete(labels = label)

#侯友宜貼文類型分析
Houtype=Su%>%group_by(Type)%>%count()
label_value <- paste('(', round(Houtype$n/sum(Houtype$n) * 100, 1), '%)', sep = '')
label=paste(Houtype$Type,label_value,sep = "")
ggplot(Houtype, aes(x="", y=n, fill=Type)) +
       geom_bar(width = 1, stat = "identity") +
       coord_polar("y", start=0)+
       labs(x = '', y = '', title = '')+
       theme(axis.text = element_blank())+
       scale_fill_discrete(labels = label)

#侯友宜6-11/24貼文數量變化
Houcount=Hou%>%group_by(month=format(Date,"%m"))%>%count()
ggplot(Houcount,aes(x=month,y=n))+
  geom_bar(stat = "identity")

#蘇貞昌6-11/24貼文數量變化
Sucount=Su%>%group_by(month=format(Date,"%m"))%>%count()
ggplot(Sucount,aes(x=month,y=n))+
  geom_bar(stat = "identity")

#蘇侯每月PO文字次數比較
Sucount=Su%>%group_by(month=format(Date,"%m"))%>%count()%>%mutate(group="Su")
Houcount=Hou%>%group_by(month=format(Date,"%m"))%>%count()%>%mutate(group="Hou")
monthcount=rbind(Sucount,Houcount)%>%arrange((month))
ggplot(monthcount,aes(x=month,y=n,fill=group))+
  geom_bar(stat="identity",position = "dodge")

ggplot(Pei,aes(x=Type,y=LIKE_COUNT))+
  geom_boxplot(col = "red")

library(ggpubr)

## Loading required package: magrittr

ggqqplot(Su$All_Reaction_Count)

library(corrplot)

## corrplot 0.84 loaded

cor(Su[c(6:14)])%>%corrplot.mixed(lower = "pie",tl.cex=0.6)

ggscatter(Su,x="All_Reaction_Count",y="LIKE_COUNT", add = "reg.line", conf.int = TRUE,cor.coef = TRUE, cor.method = "pearson")

#兩候選人like count分析
#折線圖(6到11月24)，用月作為區分
ggplot(Pei, aes(x = cDate,y = LIKE_COUNT,colour = Page_Name))+ geom_line()+
  geom_vline(xintercept = as.Date("2018-07-24"),col = "black") + 
  geom_vline(xintercept = as.Date("2018-08-24"),col = "black") +
  geom_vline(xintercept = as.Date("2018-09-24"),col = "black") +
  geom_vline(xintercept = as.Date("2018-10-24"),col = "black") +
  geom_vline(xintercept = as.Date("2018-11-24"),col = "black")