In joebrew/vilaweb: Reproducible data analysis for www.vilaweb.cat

# Basic knitr options
library(knitr)
opts_chunk$set(comment = NA, 
               echo = FALSE, 
               warning = FALSE, 
               message = FALSE, 
               error = TRUE, 
               cache = FALSE,
               fig.path = 'figures/')

# Libraries
library(vilaweb)
library(tidyverse)
library(databrew)
source('functions.R')

party_dict <- 
  tibble(P24 = c("PPC",
"CiU",
"ERC",
"PSC",
"ICV-EUiA",
"C's",
"Reagrupament.cat",
"SI",
"PxC",
"CUP",
"UPyD",
"Podemos",
"Barcelona en Comú",
"CDC",
"Junts pel Sí",
"Catalunya sí que es pot",
"Democràcia i Llibertat",
"En Comú Podem",
"PACMA",
"PDeCAT",
"Junts per Catalunya",
"Catalunya en Comú Podem",
"Altres partits",
"Cap",
"No ho sap",
"No contesta"),
partit = c("PPC",
"PDCat/CiU/CDC/Junts",
"ERC",
"PSC",
"ICV-EUiA",
"C's",
"Reagrupament.cat",
"SI",
"PxC",
"CUP",
"UPyD",
"Podem(os)",
"Podem(os)",
"PDCat/CiU/CDC/Junts",
"PDCat/CiU/CDC/Junts",
"Podem(os)",
"Democràcia i Llibertat",
"Podem(os)",
"PACMA",
"PDCat/CiU/CDC/Junts",
"PDCat/CiU/CDC/Junts",
"Podem(os)",
"Altre/Cap/NS/NC",
"Altre/Cap/NS/NC",
"Altre/Cap/NS/NC",
"Altre/Cap/NS/NC"))

pd <- vilaweb::ceo %>%
  left_join(party_dict) %>%
  group_by(partit) %>%
  mutate(size = n()) %>%
  filter(size >= 50) %>%
  ungroup %>%
  mutate(year = ANY) %>%
  mutate(axis = as.character(P25)) %>%
  mutate(axis = ifelse(axis == 'Extrema esquerra',
                       '1',
                       ifelse(axis == 'Extrema dreta',
                              '10',
                              as.character(axis)))) %>%
  mutate(axis = as.numeric(axis)) %>%
  filter(!partit %in% c('Altre/Cap/NS/NC',
                        'ICV-EUIA'))

pd <- pd %>%
  mutate(indy = P31) %>%
  mutate(indy = as.character(indy)) %>%
  mutate(indy = ifelse(indy %in% c('No ho sap',
                                  'No contesta'),
                       'NS/NC',
                       indy)) %>%
    filter(!is.na(indy)) %>%
  filter(indy != 'NS/NC') %>%
  mutate(economy = P56I) %>%
  mutate(economy = as.character(economy)) %>%
  mutate(economy = ifelse(economy %in% c('No ho sap',
                                  'No contesta'),
                       'NS/NC',
                       economy)) %>%
    filter(!is.na(economy)) %>%
  filter(economy != 'NS/NC') %>%
    mutate(economy = 
           ifelse(economy %in% c("Molt d'acord",
                                 "D'acord"),
                  "D'acord o molt d'acord",
                  ifelse(economy %in% c("En desacord",
                                        "Molt en desacord"),
                         "En desacord o molt en desacord",
                         "Ni d'acord ni en desacord"))) %>% 
  group_by(economy, partit) %>%
  summarise(n = sum(PONDERA)) %>%
  ungroup %>%
  group_by(partit) %>%
  mutate(p = n / sum(n) * 100) %>%
  mutate(`Muestra` = n)

pd$partit <- ifelse(grepl('PDCat', pd$partit),
                    'PDCat/CiU/\nCDC/Junts', pd$partit)
original_levels <- c("D'acord o molt d'acord",
                               "Ni d'acord ni en desacord",
                               "En desacord o molt en desacord")
cat_levels <- c("\nD'acord o\nmolt\nd'acord\n",
                               "\nNi d'acord\nni en\ndesacord\n",
                               "\nEn desacord\no molt en\ndesacord\n")
en_levels <- c("\nAgree or\nstrongly agree\n",
                               "\nNeither agree\nnor disagree\n",
                               "\nDisagree or\nstrongly disagree\n")
esp_levels <- c("\nDe acuerdo\no muy\nde acuerdo\n",
                               "\nNi de\nacuerdo ni\nen desacuerdo\n",
                               "\nEn desacuerdo\no muy\nen desacuerdo\n")
catalan <- FALSE
english <- TRUE
if(catalan){
  the_subtitle <- "Grau d'acord amb l'afirmació"
  the_levels <- cat_levels
  the_title <- "\n'Amb tanta immigració, un ja no se sent com a casa'"
  the_caption <- 'Mostra: 3143 residents de Catalunya amb ciutadania espanyola. Dades: Combinació enquestes CEO.\n2015 i 2018. Preguntes P56I i P31. Gràfic: Joe Brew | @joethebrew.'
} else if(english) {
  the_levels <- en_levels
  the_subtitle <- "Extent of agreement with the above phrase"
  the_title <- '\n"With so much immigration, it doesn\'t feel like home any more"'
  the_caption <- 'Sample size: 3143 residents of Catalonia with Spanish citizenship. Data: Aggregated CEO surveys,\n2015 and 2018. Questions P56I and P31. Chart: Joe Brew | @joethebrew | joe@databrew.cc.'
} else {
    the_levels <- esp_levels
  the_subtitle <- "Grado de acuerdo con la afirmación"
  the_title <- '\n"Con tanta inmigración, uno ya no se siente como en casa"'
  the_caption <- 'Muestra: 3143 residentes de Cataluña con ciudadanía española. Datos: Combinación encuestas CEO.\n2015 y 2018. Preguntas P56I y P31. Gráfico: Joe Brew | @joethebrew.'
}


pd$economy<- factor(pd$economy,
                    levels = rev(original_levels),
                    labels = rev(the_levels))
n_cols <- length(unique(pd$economy))
cols <- databrew::make_colors(n = 10, categorical = TRUE)
cols <- cols[c(2,3, 6)]
cols <- rev(cols)
# cols <- rev(cols)
cols[2] <- grey(0.2)

pd <- pd %>%
  arrange(desc(economy), p)
pd$partit <- factor(pd$partit, levels = unique(pd$partit))

ggplot(data = pd %>%
         filter(!partit %in% c("ICV-EUiA", "PACMA")),
       aes(x = partit,
           y = p)) +
  geom_bar(stat = 'identity',
           position = 'stack',
           # position = position_dodge(width = 0.9),
           # color = 'black',
           alpha = 0.9,
           aes(fill = economy)) +
  # facet_wrap(~partit) +
  theme_vilaweb() +
  # scale_fill_manual(name = '',
  #                   values = cols) +
  geom_text(aes(label = round(p, digits = 1),
                y = p-1),
            alpha = 0.9,
            col = 'white',
            vjust = 1,
            position = 'stack',
            # position = position_dodge(width = 0.9),
            size = 5) +
  scale_fill_manual(name = '',
                    values = cols) +
  labs(x = '',
       y = '%',
       subtitle = the_subtitle,
       title =  the_title,
       caption = the_caption) +
  theme(legend.position = 'right',
        axis.text.x = element_text(size = 12),
       plot.caption = element_text(hjust = 0))

pd <- vilaweb::ceo %>%
  left_join(party_dict) %>%
  group_by(partit) %>%
  mutate(size = n()) %>%
  filter(size >= 50) %>%
  ungroup %>%
  mutate(year = ANY) %>%
  mutate(axis = as.character(P25)) %>%
  mutate(axis = ifelse(axis == 'Extrema esquerra',
                       '1',
                       ifelse(axis == 'Extrema dreta',
                              '10',
                              as.character(axis)))) %>%
  mutate(axis = as.numeric(axis)) %>%
  filter(!partit %in% c('Altre/Cap/NS/NC',
                        'ICV-EUIA'))

pd <- pd %>%
  mutate(indy = P31) %>%
  mutate(indy = as.character(indy)) %>%
  mutate(indy = ifelse(indy %in% c('No ho sap',
                                  'No contesta'),
                       'NS/NC',
                       indy)) %>%
    filter(!is.na(indy)) %>%
  filter(indy != 'NS/NC') %>%
  mutate(economy = P56I) %>%
  mutate(economy = as.character(economy)) %>%
  mutate(economy = ifelse(economy %in% c('No ho sap',
                                  'No contesta'),
                       'NS/NC',
                       economy)) %>%
    filter(!is.na(economy)) %>%
  filter(economy != 'NS/NC') %>%
    mutate(economy = 
           ifelse(economy %in% c("Molt d'acord",
                                 "D'acord"),
                  "D'acord o molt d'acord",
                  ifelse(economy %in% c("En desacord",
                                        "Molt en desacord"),
                         "En desacord o molt en desacord",
                         "Ni d'acord ni en desacord"))) %>% 
  group_by(economy, indy) %>%
  summarise(n = sum(PONDERA)) %>%
  ungroup %>%
  group_by(indy) %>%
  mutate(p = n / sum(n) * 100) %>%
  mutate(`Muestra` = n)

original_levels <- c("D'acord o molt d'acord",
                               "Ni d'acord ni en desacord",
                               "En desacord o molt en desacord")
cat_levels <- c("\nD'acord o\nmolt\nd'acord\n",
                               "\nNi d'acord\nni en\ndesacord\n",
                               "\nEn desacord\no molt en\ndesacord\n")
en_levels <- c("\nAgree or\nstrongly agree\n",
                               "\nNeither agree\nnor disagree\n",
                               "\nDisagree or\nstrongly disagree\n")
esp_levels <- c("\nDe acuerdo\no muy\nde acuerdo\n",
                               "\nNi de\nacuerdo ni\nen desacuerdo\n",
                               "\nEn desacuerdo\no muy\nen desacuerdo\n")
catalan <- FALSE
english <- TRUE



if(catalan){
  the_subtitle <- "Grau d'acord amb l'afirmació"
  the_levels <- cat_levels
  the_title <- "\n'Amb tanta immigració, un ja no se sent com a casa'"
  the_caption <- 'Mostra: 3143 residents de Catalunya amb ciutadania espanyola. Dades: Combinació enquestes CEO.\n2015 i 2018. Preguntes P56I i P31. Gràfic: Joe Brew | @joethebrew.'
    indies <- c('A favor\nde la\nindependència',
              'NS/NC',
              'En contra\nde la\nindependència')
} else if(english){
  the_levels <- en_levels
  the_subtitle <- "Extent of agreement with the phrase"
  the_title <- '\n"With so much immigration, it doesn\'t feel like home any more"'
  the_caption <- 'Sample size: 3143 residents of Catalonia with Spanish citizenship. Data: Aggregated CEO surveys,\n2015 and 2018. Questions P56I and P31. Chart: Joe Brew | @joethebrew | joe@databrew.cc\nCode for reproducibility: https://github.com/joebrew/vilaweb/tree/master/analyses/xenofobia.'
  indies <- c('In favor of\nCatalan independence',
              'Not sure/no answer',
              'Against\nCatalan independence')
} else {
    the_levels <- esp_levels
  the_subtitle <- "Grado de acuerdo con la afirmación"
  the_title <- '\n"Con tanta inmigración, uno ya no se siente como en casa"'
  the_caption <- 'Muestra: 3143 residentes de Cataluña con ciudadanía española. Datos: Combinación encuestas CEO.\n2015 y 2018. Preguntas P56I y P31. Gráfico: Joe Brew | @joethebrew.'
  indies <- c('A favor\nde la\nindependencia',
              'NS/NC',
              'En contra\nde la\nindependencia')
}

pd$indy <- 
  ifelse(pd$indy == 'Sí', indies[1],
         ifelse(pd$indy == 'No', indies[3],
                indies[2]))

pd$economy<- factor(pd$economy,
                    levels = rev(original_levels),
                    labels = rev(the_levels))
n_cols <- length(unique(pd$economy))
cols <- databrew::make_colors(n = 10, categorical = TRUE)
cols <- cols[c(2,3, 6)]
cols <- rev(cols)
# cols <- rev(cols)
cols[2] <- grey(0.2)

pd <- pd %>%
  arrange(desc(economy), p)
pd$indy <- factor(pd$indy, levels = unique(pd$indy))

ggplot(data = pd,
       aes(x = indy,
           y = p)) +
  geom_bar(stat = 'identity',
           position = 'stack',
           # position = position_dodge(width = 0.9),
           # color = 'black',
           alpha = 0.85,
           aes(fill = economy)) +
  # facet_wrap(~partit) +
  theme_vilaweb() +
  # scale_fill_manual(name = '',
  #                   values = cols) +
  geom_text(aes(label = paste0(round(p, digits = 1), '%'),
                y = p-1),
            alpha = 0.9,
            col = 'white',
            vjust = 1,
            position = 'stack',
            # position = position_dodge(width = 0.9),
            size = 5) +
  scale_fill_manual(name = '',
                    values = cols) +
  labs(x = '',
       y = '%',
       subtitle = the_subtitle,
       title =  the_title,
       caption = the_caption) +
  theme(legend.position = 'right',
        axis.text.x = element_text(size = 12),
       plot.caption = element_text(hjust = 0,
                                   color = cols[1]))