global.R

library(dplyr)
library(wordcloud)
library(RColorBrewer)
library(shinythemes)

device_data=read.csv('./data/phonedata.csv', header=T, stringsAsFactors = F)
#str(device_data)
device_map=dplyr::filter(device_data, !is.na(longitude), !is.na(latitude), !is.na(group))%>%
  dplyr::filter(longitude>=73, longitude<136, latitude>=4, latitude<54)
#str(device_map)
top10 <- names(sort(table(device_map$phone_brand_English), decreasing = T))[1:10]
gender=unique(device_map$gender)
agegroup=c("F23-", "F24-26", "F27-28", "F29-32", "F33-42", "F43+", "M22-", "M23-26", "M27-28", "M29-31", "M32-38", "M39+")

table1=device_map%>%
  dplyr::group_by(phone_brand_English)%>%
  dplyr::summarise(n=n())%>%
  arrange(desc(n))

phoneprice=read.csv('./data/phoneprice.csv',header=T, stringsAsFactors = F)
phoneprice=head(phoneprice,4)

device_map%>%
  dplyr::group_by(phone_brand_English, gender) %>%
  dplyr::summarise(n=n()) %>%
  mutate(percent=n/sum(n))-> phone_bygender

# # further analysi of shiny app
# phone_male=filter(phone_bygender, gender=='M')
# sum(phone_male$n)/sum(phone_bygender$n)
# 
# N=nrow(device_map)
# agegroup=rep(0, N)
# for (i in 1:N){
#   if (device_map$age[i]<=26){
#     agegroup[i]='post-90s'
#   }
#   else if (device_map$age[i]>26&device_map$age[i]<=36){
#     agegroup[i]='post-80s'
#   }
#   else if (device_map$age[i]>36&device_map$age[i]<=46){
#     agegroup[i]='post-70s'
#   }
#   else if (device_map$age[i]>46&device_map$age[i]<=56){
#     agegroup[i]='post-60s'
#   }
#   else if (device_map$age[i]>56){
#     agegroup[i]='post-50s'
#   }
# }
# device_map=mutate(device_map, agegroup=agegroup)
# 
# agedis=device_map%>%
#   dplyr::filter(phone_brand_English %in% c('Xiaomi', 'Huawei', 'OPPO', 
#                                            'vivo', 'samsung')) %>%
#   dplyr::group_by(agegroup, phone_brand_English)%>%
#   summarise(n=n())%>%
#   mutate(percent=n/sum(n))
# ageplot=ggplot(data=agedis, aes(x=agegroup, y=percent,
#                         fill=phone_brand_English))+
#   geom_bar(stat = 'identity')+
#   xlab('age group')+
#   ylab('percent of number')+
#   ggtitle('User Age Group Distribution of Top 5 Phone Brands')+
#   theme_bw()
# ggplotly(ageplot)



app_data=read.csv('./data/appmap.csv', header=T, stringsAsFactors = F)
#str(app_data)

app_map=dplyr::filter(app_data, !is.na(longitude), !is.na(latitude), !is.na(group))%>%
  dplyr::filter(longitude>=73, longitude<136, latitude>=4, latitude<54)
#str(app_map)

top10APP <- names(sort(table(app_map$category), decreasing = T))[1:10]

# # further analysi of shiny app
# N1=nrow(app_map)
# agegroup1=rep(0, N1)
# for (i in 1:N1){
#   if (app_map$age[i]<=26){
#     agegroup1[i]='post-90s'
#   }
#   else if (app_map$age[i]>26&app_map$age[i]<=36){
#     agegroup1[i]='post-80s'
#   }
#   else if (app_map$age[i]>36&app_map$age[i]<=46){
#     agegroup1[i]='post-70s'
#   }
#   else if (app_map$age[i]>46&app_map$age[i]<=56){
#     agegroup1[i]='post-60s'
#   }
#   else if (app_map$age[i]>56){
#     agegroup1[i]='post-50s'
#   }
# }
# app_map=mutate(app_map, agegroup=agegroup1)
# unique(app_map$category)
# 
# agedis1=app_map%>%
#   dplyr::filter(category %in% c("Property Industry 2.0", "Industry tag" , "video" , 
#                                  "Services 1" ,"P2P net loan" ) ) %>%
#   dplyr::group_by(agegroup, category)%>%
#   summarise(n=n())%>%
#   mutate(percent=n/sum(n))
# ageplot1=ggplot(data=agedis1, aes(x=agegroup, y=percent,
#                                 fill=category))+
#   geom_bar(stat = 'identity')+
#   xlab('age group')+
#   ylab('percent of number')+
#   ggtitle('User Age Group Distribution of APP cateogoreis')+
#   theme_bw()
# ggplotly(ageplot1)



table2=app_map%>%
  dplyr::group_by(category)%>%
  dplyr::summarise(n=n())%>%
  arrange(desc(n))

app_map%>%
  dplyr::group_by(category, gender) %>%
  dplyr::summarise(n=n()) %>%
  mutate(percent=n/sum(n))-> app_bygender

app_map%>%
  dplyr::group_by(category, phone_brand_English, gender, age) %>%
  dplyr::summarise(n=n()) -> app_byphone

app_map$is_active=as.character(app_map$is_active)
app_map %>%
dplyr::group_by(category, is_active) %>%
  dplyr::summarise(n=n())->app_byactive

app_map%>%
  dplyr::group_by(category, phone_brand_English)%>%
  dplyr::summarise(n=n())->app_bybrand