####    S# ：数据编号
####    Location：案件发生城市，州
####    Date：时间
####    Summary：案件总结
####    Fatalities：死亡人数
####    Injured：受伤人数
####    Total victims：受害者总人数
####    Mental Health Issues：精神状况
####    Race：种族
####    Gender：性别
####    Latitude：纬度
####    Longitude：经度

library(tidyverse)
library(stringr)
library(data.table)
library(maps)
library(lubridate)
library(leaflet)

shooting <- read.csv('Mass Shootings Dataset Ver 2.csv',stringsAsFactors = F,header = T)
summary(shooting)
glimpse(shooting)

# 将Date字段进行转化,同时创建新的变量year
shooting <- shooting %>% select(1:13) %>% mutate(Date=mdy(shooting$Date),year=year(Date))
summary(shooting$year)# 对性别进行提取
shooting$Gender<-if_else(shooting$Gender=="M","Male",shooting$Gender)
# 对种族字段进行提取
shooting$Race<-if_else(str_detect(shooting$Race,"Black American or African American"),"Black",shooting$Race)
shooting$Race<-if_else(str_detect(shooting$Race,"White American or European American"),"White",shooting$Race)
shooting$Race<-if_else(str_detect(shooting$Race,"Asian American"),"Asian",shooting$Race)
shooting$Race<-if_else(str_detect(shooting$Race,"Some other race"),"Other",shooting$Race)
shooting$Race<-if_else(str_detect(shooting$Race,"Native American or Alaska Native"),"Native American",shooting$Race)
# 对时间数据进行切分
shooting$yearcut<-cut(shooting$year,breaks = 10)
# 对是否有心理疾病进行处理
shooting$Mental.Health.Issues<-if_else(str_detect(shooting$Mental.Health.Issues,"Un"),"Unknown",shooting$Mental.Health.Issues)
shooting$Race<-str_to_upper(shooting$Race)
shooting$Mental.Health.Issues<-str_to_upper(shooting$Mental.Health.Issues)
# 把location分解成city和state两个变量
shooting$city <- sapply(shooting$Location,function(x){return(unlist(str_split(x,','))[1] %>% str_trim())
})shooting$state <- sapply(shooting$Location,function(x){return(unlist(str_split(x,','))[2] %>% str_trim())
})

# 每年受到枪击的死亡人数
shooting %>% group_by(year) %>%summarise(total=sum(Total.victims)) %>%ggplot(aes(x=year,y=total)) +geom_bar(stat = 'identity',fill='blue') +geom_text(aes(label=total),vjust=-0.2) +xlim(1969,2020) +geom_line(color='red') +ylab('Total victims every year') +ggtitle('People died because of gun shoot every year')

# 受伤人数的地理位置分布
shooting %>%select(Total.victims,Fatalities,Longitude,Latitude,Summary) %>%na.omit() %>%leaflet() %>%addProviderTiles(providers$OpenStreetMap) %>%fitBounds(-124,30,-66,43) %>%addCircles(color='#8A0707',lng = ~Longitude,lat = ~Latitude,weight = 1,radius = ~sqrt(Total.victims) * 20000,popup = ~Summary)# 死亡人数的地理位置分布
shooting %>%select(Total.victims,Fatalities,Longitude,Latitude,Summary) %>%na.omit() %>%leaflet() %>%addProviderTiles(providers$OpenStreetMap) %>%fitBounds(-124,30,-66,43) %>%addCircles(color='blue',lng = ~Longitude,lat = ~Latitude,weight = 1,radius = ~sqrt(Fatalities) * 20000,popup = ~Summary)

shooting %>%ggplot(aes(x=factor(Gender),fill=factor(Gender)))+geom_bar()+xlab('Gender')+ylab('Number of each Gender')+ggtitle('The distribution of gender')

shooting %>% na.omit() %>%group_by(Race) %>%summarise(num=sum(Total.victims)) %>%ggplot(aes(x=factor(Race),y=num,fill=factor(Race)))+geom_bar(stat = 'identity')+coord_polar(theta = 'y')+labs(x='Race',y='Number of killed people',fill='Race')+ggtitle('People killed by different race')

shooting %>%mutate(month=month(Date)) %>%group_by(month) %>%summarise(n=sum(Total.victims)) %>%ggplot(aes(x=factor(month),y=n)) +geom_bar(stat = 'identity')+labs(x='month',y='Number of killed people')+ggtitle('The distribution of killed people every month')+geom_text(aes(label=n),vjust=-0.2,color='red')+theme_bw()

shooting %>% na.omit() %>% ggplot(aes(x=Mental.Health.Issues)) + geom_bar()+scale_x_discrete(limits=c("NO","YES"))+theme_bw()

shooting %>%na.omit() %>%group_by(Mental.Health.Issues) %>%summarise(n=sum(Total.victims)) %>%ggplot(aes(x=factor(Mental.Health.Issues),y=n,group=1)) +geom_bar(stat = 'identity',fill='pink')+scale_x_discrete(limits=c('NO','YES'))+geom_text(aes(label=n),vjust=-0.2)+geom_line(color='red')

shooting %>%na.omit() %>%group_by(yearcut) %>%ggplot(aes(x=yearcut,fill=Race))+geom_bar(position = 'dodge')

# 通过正则表达式从摘要中提取年龄
tmp <- mutate(shooting,age=str_extract_all(shooting$Summary,pattern = '(,\\s)\\d{2}(,)'),age2 = str_extract_all(shooting$Summary,pattern = '(a\\s)\\d{2}(-year)'))
tmp$age <- str_sub(tmp$age,3,4)
tmp$age2 <- str_sub(tmp$age2,3,4)
# 去掉年龄不明的字段
te <- subset(tmp,tmp$age != 'ar')
te2 <- subset(tmp,tmp$age2 != 'ar')
te <- rbind(te,te2)for(i in 1:nrow(te)){if(te$age[i] == 'ar'){te$age[i] = te$age2[i]}
}
te <- arrange(te,age)
te <- te[-c(1:4),]
te <- arrange(te,S.)
te$age <- as.integer(te$age)
te3 <- te %>%select(S.,age) %>%mutate(agecut=cut(te$age,breaks = 10*(1:7)))
shoot_age <- left_join(te3,shooting)

ggplot(data=shoot_age,aes(x=agecut))+geom_bar(fill='blue')+theme_bw()

ggplot(data=shoot_age,aes(x=agecut,fill=Mental.Health.Issues))+geom_bar()

# 城市分布
shooting %>%group_by(city) %>%summarise(count=n()) %>%filter(city != '' & count >= 2) %>%ggplot(aes(x=reorder(city,count),y=count))+geom_bar(stat = 'identity',fill='lightblue')+coord_flip()+labs(x='City',y='Number of gun-shot happended')+ggtitle('The number of case happened in each city')# 州分布
shooting %>%group_by(state) %>%summarise(count=n()) %>%filter(state != '' & count >= 2) %>%ggplot(aes(reorder(state,count),y=count))+geom_bar(stat='identity',fill='lightblue')+coord_flip()+labs(x='State',y='Number of gun-shot happended')+ggtitle('The number of case happened in each state')