library(tidyverse)
library(lubridate)
library(extrafont)
loadfonts()

## 맥 인코딩 ----- 
# Sys.setlocale('LC_CTYPE', 'ko_KR.UTF-8')
options(encoding = "utf8")

## 데이터 3구 연도별 종합 -----
sungnam_2016_df <- read_rds("data/sungnam_2016_df.rds") %>% mutate(`연도` = ymd("2016-03-31") %>% year(.))
sungnam_2017_df <- read_rds("data/sungnam_2017_df.rds") %>% mutate(`연도` = ymd("2017-03-31") %>% year(.))
sungnam_2018_df <- read_rds("data/sungnam_2018_df.rds") %>% mutate(`연도` = ymd("2018-03-31") %>% year(.))
sungnam_2019_df <- read_rds("data/sungnam_2019_df.rds") %>% mutate(`연도` = ymd("2019-02-28") %>% year(.))

sungnam_df <- bind_rows(sungnam_2016_df, sungnam_2017_df) %>% 
  bind_rows(sungnam_2018_df) %>% 
  bind_rows(sungnam_2019_df) %>% 
  mutate(`연령` = as.integer(`연령`))

## 성남 3구 연령대 데이터프레임 -----
sungnam_age_df <- sungnam_df %>% 
  mutate(`연령대` = case_when(`연령` >=0 & `연령` <=5 ~   "0-5세",
                              `연령` >=6 &  `연령` <=10 ~ "06-10세",
                              `연령` >=11 & `연령` <=15 ~ "11-15세",
                              `연령` >=16 & `연령` <=20 ~ "16-20세",
                              `연령` >=21 & `연령` <=25 ~ "21-25세",
                              `연령` >=26 & `연령` <=30 ~ "26-30세",
                              `연령` >=31 & `연령` <=35 ~ "31-35세",
                              `연령` >=36 & `연령` <=40 ~ "36-40세",
                              `연령` >=41 & `연령` <=45 ~ "41-45세",
                              `연령` >=46 & `연령` <=50 ~ "46-50세",
                              `연령` >=51 & `연령` <=55 ~ "51-55세",
                              `연령` >=56 & `연령` <=60 ~ "56-60세",
                              `연령` >=61 & `연령` <=65 ~ "61-65세",
                              `연령` >=66 & `연령` <=70 ~ "66-70세",
                              `연령` >=71 & `연령` <=75 ~ "71-75세",
                              `연령` >=76 & `연령` <=80 ~ "76-80세",
                              `연령` >=81  ~ "80세+")) %>% 
  mutate(`연령대` = factor(`연령대`,levels=c("0-5세","06-10세","11-15세","16-20세","21-25세","26-30세","31-35세","36-40세","41-45세","46-50세","51-55세","56-60세","61-65세","66-70세","71-75세","76-80세","80세+")))  %>% 
    mutate(`선거구` = case_when(`행정동` %in%  c("구미1동", "구미동", "금곡동", "분당동", "수내1동", "수내2동", "수내3동","정자1동", 
"정자2동", "정자3동", "정자동") ~ "분당을",
                             TRUE ~ "분당갑"))


## 분당을 연령대 데이터 -----

bundang_eul_df <- sungnam_age_df %>% 
  filter(`선거구` == "분당을") %>% 
  group_by(`연도`, `연령대`, `행정동`) %>% 
  summarise(`인구수` = sum(`인구수`)) %>% 
  ungroup() %>% 
  mutate(`연도` = factor(`연도`))

bundang_eul_df %>% write_csv("data/bundang_eul_df.csv")

library(tidyverse)
library(trelliscopejs)
library(rbokeh)

bundang_eul_df <- read_csv("data/bundang_eul_df.csv") 

bundang_eul_df <- bundang_eul_df %>% 
  mutate(`연령대` = factor(`연령대`, levels=c("0-5세","06-10세","11-15세","16-20세","21-25세","26-30세","31-35세","36-40세","41-45세","46-50세","51-55세","56-60세","61-65세","66-70세","71-75세","76-80세","80세+")))

bundang_eul_df %>% 
  filter(연도 %in% c(2016, 2019),
         행정동 == "구미동") %>% 
  mutate(연도 = as.factor(연도)) %>% 
  figure() %>% 
    ly_bar(x = 연령대, y = 인구수, color = 연도, 
           position = "dodge",
           hover = TRUE) %>% 
    theme_axis("x", major_label_orientation = 90) %>% 
    set_palette(discrete_color = pal_color(c("darkgray", "blue")))

bundang_eul_comp_df <- bundang_eul_df %>% 
  filter(연도 %in% c(2016, 2019)) %>% 
  mutate(연도 = as.factor(연도))

bundang_eul_list <- split(bundang_eul_comp_df, bundang_eul_comp_df$행정동)

## 시각화 함수
age_plot <- function(df) {
  df %>% 
  figure(legend_location = NULL,
         xlab="", ylab="") %>% 
    ly_bar(x = 연령대, y = 인구수, color = 연도, 
           position = "dodge",
           hover = TRUE) %>% 
    theme_axis("x", major_label_orientation = 90) %>% 
    set_palette(discrete_color = pal_color(c("darkgray", "blue")))
  }

## 데이터를 함수에 적합(apply)
bundang_fig_list <- lapply(bundang_eul_list, age_plot)

## bokeh 시각화
grid_plot(bundang_fig_list, same_axes = TRUE, 
          width = 800, height=600, 
          nrow = 3)

options(encoding = "utf-8")

# 연도별, 행정동 중첩시킴.
by_year_dong <- bundang_eul_df %>%
  filter(연도 %in% c(2016, 2019)) %>% 
  dplyr::mutate(연도 = factor(연도)) %>%
  group_by(`행정동`) %>%
  nest()

# 시각화
by_year_dong_bokeh <- by_year_dong %>% mutate(
  panel = map_plot(data,
    ~ .x %>% 
      figure(legend_location = NULL,
             xlab="", ylab="",
             title="분당을 동별 2016 vs 2019년 인구구조 변화") %>% 
      ly_bar(`연령대`, `인구수`, color=`연도`,  position = "dodge", hover = TRUE) %>% 
      theme_axis("x", major_label_orientation = 90) %>% 
      set_palette(discrete_color = pal_color(c("darkgray", "blue")))
  )) 
  
by_year_dong_bokeh %>%  
  trelliscope(name="bundang_viz", nrow = 1, ncol = 2, path = "bundang_viz_files")