analyses/left_right_party/functions.R

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

commafy <- function(x){
  scales::comma(x, big.mark = '.', decimal.mark = ',')
}
point_replace <- function(x){
  gsub(',', '.', x, fixed = TRUE)
}


# Functions
mround <- function(x,base){ 
  base*round(x/base) 
} 

round_percent <- function(x) { 
  x <- x/sum(x)*100  # Standardize result
  res <- floor(x)    # Find integer bits
  rsum <- sum(res)   # Find out how much we are missing
  if(rsum<100) { 
    # Distribute points based on remainders and a random tie breaker
    o <- order(x%%1, sample(length(x)), decreasing=TRUE) 
    res[o[1:(100-rsum)]] <- res[o[1:(100-rsum)]]+1
  } 
  res 
}

# # Read in data manually (until package built)
# cis_2018_09 <- haven::read_sav('../../../data-raw/cis/monthly/2018-09/3223.sav')
# cis_2018_09 <- haven::as_factor(cis_2018_09)
# cis_list <- list('2018-09' = cis_2018_09)

overall_plot <- function(language = 'en',
                         geo = c('cat', 'esp', 'all'),
                         return_table = FALSE){
  
  cis_data  <- cis_list$`2018-09`
  
  if(length(geo) != 1){
    stop('geo must be one of "cat", "esp", "all"')
  }
  
  if(geo == 'esp'){
    plot_data <- cis_data %>% filter(CCAA != 'Cataluña')  
    if(language == 'en'){subtitle <- 'Spain, not counting Catalonia'} else {subtitle <-'Espanya, sense incloure Catalunya'}
  }
  if(geo == 'cat'){
    plot_data <- cis_data %>% filter(CCAA == 'Cataluña')
    if(language == 'en'){subtitle <- 'Catalonia'} else {subtitle <-'Catalunya'}
  }
  if(geo == 'all'){
    if(language == 'en'){subtitle <- 'Spain, including Catalonia'} else {subtitle <- 'Espanya, incloent Catalonia'}
    plot_data <- cis_data 
  }
  
  
  plot_data <- plot_data %>%
    mutate(P13 = as.character(P13)) %>%
    # Combine the no answer / not sures, etc.
    mutate(P13 = ifelse(P13 == 'N.C.',
                        '(NO LEER) Está en duda, no sabe lo suficiente',
                        P13)) %>%
    group_by(es = P13) %>%
    tally %>%
    ungroup %>%
    mutate(es = as.character(es)) %>%
    mutate(y  = n / sum(n) * 100)
  right <- 
    data_frame(
      # Manually copied from the Sep 2018 CIS avance de resultados
      # y = c(7.4, 39.9, 17.5, 19.5, 8.2, 7.4),
      es = c('Muy satisfecho/a',
             'Bastante satisfecho/a',
             '(NO LEER) Regular',
             'Poco satisfecho/a',
             'Nada satisfecho/a',
             '(NO LEER) Está en duda, no sabe lo suficiente'),
      ca = factor(c('Molt satisfet',
                    'Bastant satisfet',
                    'Regular',
                    'Poc satisfet',
                    'Gens satisfet',
                    'NS/NC'),
                  levels = c('Molt satisfet',
                             'Bastant satisfet',
                             'Regular',
                             'NS/NC',
                             'Poc satisfet',
                             'Gens satisfet')),
      en = factor(c('Very satisfied',
                    'Satisfied',
                    'So-so',
                    'Not satisfied',
                    'Not at all satisfied',
                    'No answer'),
                  levels = c('Very satisfied',
                             'Satisfied',
                             'So-so',
                             'No answer',
                             'Not satisfied',
                             'Not at all satisfied')))
  plot_data <- left_join(plot_data, right, by = 'es')
  if(language == 'en'){
    title <- 'Level of satisfaction with Spanish Constitution'
    caption <- 'Data from CIS survey, September 2018.\nJoe Brew | @joethebrew.'
    plot_data$x <- plot_data$en
  } else {
    title <- 'Nivell de satisfacció amb la Constitució Espanyola'
    caption <- 'Dades del CIS, Setembre 2018.\nJoe Brew | @joethebrew.'
    plot_data$x <- plot_data$ca
  }
  
  bp <- function(x,z){RColorBrewer::brewer.pal(n = 8,name =x)[z]}
  cols <- c(bp('Blues', 8),
            bp('Blues', 5),
            bp('Greys', 3),
            bp('Greys', 6),
            bp('Oranges', 5),
            bp('Oranges', 8))
  if(return_table){
    return(plot_data)
  }
  
  plot_data <- plot_data %>% arrange(x)
  x <- plot_data$y
  names(x) <- plot_data$x
  x <- round_percent(x)
  names(x) <- factor(levels(plot_data$x))
  
  gg <- waffle(parts = x,
               rows = 10,
               colors = cols) +
    theme_vilaweb() +
    theme(legend.position = 'right') +
    scale_x_continuous(expand = c(0, 0)) +
    scale_y_continuous(expand = c(0, 0)) +
    coord_equal() +
    theme(panel.grid = element_blank()) +
    theme(panel.border = element_blank()) +
    theme(panel.background = element_blank()) +
    theme(panel.spacing = unit(0, "null")) +
    theme(axis.text = element_blank()) +
    theme(axis.title.x = element_text(size = 10)) +
    theme(axis.ticks = element_blank()) +
    theme(axis.line = element_blank()) +
    theme(axis.ticks.length = unit(0, "null")) +
    theme(plot.title = element_text(size = 18)) +
    theme(plot.background = element_blank()) +
    theme(panel.spacing = unit(c(0, 0, 0, 0), "null")) +
    guides(fill=guide_legend(ncol=1)) + 
    theme(legend.text = element_text(size = 14),
          plot.subtitle = element_text(size = 16))  +
    labs(title = title,
         subtitle = subtitle,
         caption = caption) 
  
  return(gg)
  # sample size 2972
}

comparison_plot <- function(language = 'en',
                            return_table = FALSE){
  
  plot_data  <- cis_list$`2018-09`
  
  
  plot_data <- plot_data %>%
    mutate(P13 = as.character(P13)) %>%
    # Combine the no answer / not sures, etc.
    mutate(P13 = ifelse(P13 == 'N.C.',
                        '(NO LEER) Está en duda, no sabe lo suficiente',
                        P13)) %>%
    group_by(es = P13, ccaa = CCAA) %>%
    tally %>%
    ungroup %>%
    mutate(es = as.character(es)) %>%
    group_by(ccaa) %>%
    mutate(y  = n / sum(n) * 100) %>%
    ungroup 
  
  right <- 
    data_frame(
      # Manually copied from the Sep 2018 CIS avance de resultados
      # y = c(7.4, 39.9, 17.5, 19.5, 8.2, 7.4),
      es = c('Muy satisfecho/a',
             'Bastante satisfecho/a',
             '(NO LEER) Regular',
             'Poco satisfecho/a',
             'Nada satisfecho/a',
             '(NO LEER) Está en duda, no sabe lo suficiente'),
      ca = factor(c('Molt satisfet',
                    'Bastant satisfet',
                    'Regular',
                    'Poc satisfet',
                    'Gens satisfet',
                    'NS/NC'),
                  levels = c('Molt satisfet',
                             'Bastant satisfet',
                             'Regular',
                             'NS/NC',
                             'Poc satisfet',
                             'Gens satisfet')),
      en = factor(c('Very satisfied',
                    'Satisfied',
                    'So-so',
                    'Not satisfied',
                    'Not at all satisfied',
                    'No answer'),
                  levels = c('Very satisfied',
                             'Satisfied',
                             'So-so',
                             'No answer',
                             'Not satisfied',
                             'Not at all satisfied')))
  plot_data <- left_join(plot_data, right, by = 'es')
  
  
  aa <- plot_data %>%
    group_by(ccaa) %>%
    summarise(satisfied = sum(y[en %in% c('Very satisfied',
                                          'Satisfied')]),
              unsatisfied = sum(y[en %in% c('Not satisfied',
                                            'Not at all satisfied')])) %>%
    mutate(bad = unsatisfied > satisfied)
  
  x <- 'CCAA'
  y <- 'Percentage'
  if(language == 'en'){
    subtitle <- 'By CCAA (autonomous community)'
    xxx <- 'Not at all satisfied'
    title <- 'Level of satisfacation with Spanish Constitution'
    caption <- 'Data from CIS survey, September 2018.\nChart: Joe Brew | @joethebrew.'
    plot_data$x <- plot_data$en
  } else {
    subtitle <- 'Per comunitat autònoma'
    xxx <- 'Gens satisfet'
    title <- 'Nivell de satisfacció amb la Constitució Espanyola'
    caption <- 'Dades del CIS, Setembre 2018.\nElaboració del gràfic: Joe Brew | @joethebrew.'
    plot_data$x <- plot_data$ca
  }
  
  bp <- function(x,z){RColorBrewer::brewer.pal(n = 8,name =x)[z]}
  cols <- c(bp('Blues', 8),
            bp('Blues', 5),
            bp('Greys', 3),
            bp('Greys', 6),
            bp('Oranges', 5),
            bp('Oranges', 8))
  if(return_table){
    return(plot_data)
  }
  
  plot_data <- plot_data %>% 
    filter(!grepl('Ceuta', ccaa),
           !grepl('Melilla', ccaa)) %>%
    mutate(ccaa = gsub(' (Principado de)', '', ccaa, fixed = TRUE),
           ccaa = gsub(' (Comunidad de)', '', ccaa, fixed = TRUE),
           ccaa = gsub(' (Región de)', '', ccaa, fixed = TRUE),
           ccaa = gsub(' (Comunidad Foral de)', '', ccaa, fixed = TRUE),
           ccaa = gsub('unitat', '.', ccaa, fixed = TRUE))
  plot_data <- plot_data %>% group_by(ccaa) %>% mutate(rangey = max(y[x == xxx], rm.na = TRUE)) %>% arrange(desc(rangey))
  plot_data$ccaa <- factor(plot_data$ccaa,
                           levels = unique(plot_data$ccaa))
  ggplot(data = plot_data,
         aes(x = ccaa,
             y = y,
             fill = x)) +
    geom_bar(stat = 'identity',
             position = position_stack(),
             width = 0.9) +
    scale_fill_manual(name = '',
                      values = cols) +
    theme_vilaweb() +
    theme(axis.text.x = element_text(angle = 90, hjust = 1,
                                     vjust = 0.5),
          legend.text = element_text(size = 13)) +
    theme(legend.position = 'right') +
    guides(fill=guide_legend(ncol=1)) +
    geom_hline(yintercept = 50,
               lty = 2,
               alpha = 0.5) +
    labs(x = x,
         y = y,
         title = title,
         subtitle = subtitle,
         caption = caption)
}

ideology_spain_plot <- function(language = 'en',
                                geo = 'esp',
                                return_table = FALSE){
  
  plot_data  <- cis_list$`2018-09`
  
  if(geo == 'esp'){
    plot_data <- plot_data %>%
      filter(CCAA != 'Cataluña') 
  } else if(geo == 'cat'){
    plot_data <- plot_data %>%
      filter(CCAA == 'Cataluña') 
  }
  plot_data <- plot_data %>%
    mutate(P13 = as.character(P13)) %>%
    # Combine the no answer / not sures, etc.
    mutate(P13 = ifelse(P13 == 'N.C.',
                        '(NO LEER) Está en duda, no sabe lo suficiente',
                        P13)) %>%
    mutate(axis = P23) %>%
    mutate(axis = ifelse(axis %in% c('N.S.', 'N.C.'),
                         'NS/NC',
                         ifelse(axis %in% '1 Izquierda', as.character('01'),
                                ifelse(axis %in% '10 Derecha', as.character(10),
                                       as.character(axis))))) %>%
    mutate(axis = as.numeric(axis))
  
  # Join for the axis translations
  right <- data_frame(axis = 1:10,
                      ideology_en = factor(c(rep('Left', 3), rep('Center', 4), rep('Right', 3)),
                                           levels = c('Left', 'Center', 'Right')),
                      ideology_ca = factor(c(rep('Esquerra', 3), rep('Centre', 4), rep('Dreta', 3)),
                                           levels = c('Esquerra', 'Centre', 'Dreta')))
  plot_data <- left_join(plot_data, right, by = 'axis')
  if(language == 'en'){
    plot_data$new_axis <- plot_data$ideology_en
  } else {
    plot_data$new_axis <- plot_data$ideology_ca
  }
  plot_data$axis <- plot_data$new_axis
  
  plot_data <- plot_data %>%
    group_by(es = P13, 
             # ccaa = ifelse(CCAA == 'Cataluña', 'CAT', 'ESP'),
             axis) %>%
    tally %>%
    ungroup %>%
    mutate(es = as.character(es)) %>%
    group_by(#ccaa,
      axis) %>%
    mutate(y  = n / sum(n) * 100) %>%
    ungroup 
  right <- 
    data_frame(
      # Manually copied from the Sep 2018 CIS avance de resultados
      # y = c(7.4, 39.9, 17.5, 19.5, 8.2, 7.4),
      es = c('Muy satisfecho/a',
             'Bastante satisfecho/a',
             '(NO LEER) Regular',
             'Poco satisfecho/a',
             'Nada satisfecho/a',
             '(NO LEER) Está en duda, no sabe lo suficiente'),
      ca = factor(c('Molt satisfet',
                    'Bastant satisfet',
                    'Regular',
                    'Poc satisfet',
                    'Gens satisfet',
                    'NS/NC'),
                  levels = c('Molt satisfet',
                             'Bastant satisfet',
                             'Regular',
                             'NS/NC',
                             'Poc satisfet',
                             'Gens satisfet')),
      en = factor(c('Very satisfied',
                    'Satisfied',
                    'So-so',
                    'Not satisfied',
                    'Not at all satisfied',
                    'No answer'),
                  levels = c('Very satisfied',
                             'Satisfied',
                             'So-so',
                             'No answer',
                             'Not satisfied',
                             'Not at all satisfied')))
  plot_data <- left_join(plot_data, right, by = 'es')
  
  
  
  x <- 'CCAA'
  y <- 'Percentage'
  if(language == 'en'){
    subtitle <- 'By ideology. All of Spain, not including Catalonia.'
    xxx <- 'Not at all satisfied'
    title <- 'Level of satisfacation with Spanish Constitution'
    caption <- 'Data from CIS survey, September 2018.\nChart: Joe Brew | @joethebrew.\nIdeology: 1-3; left; 4-7: center; 8-10: right.'
    plot_data$x <- plot_data$en
  } else {
    subtitle <- 'Per ideologia. Tota Espanya, sense incloure Catalunya.'
    xxx <- 'Gens satisfet'
    title <- 'Nivell de satisfacció amb la Constitució Espanyola'
    caption <- 'Dades del CIS, Setembre 2018.\nElaboració del gràfic: Joe Brew | @joethebrew.\nIdeologia: 1-3: esquerra; 4-7: centre; 8-10: dreta.'
    plot_data$x <- plot_data$ca
  }
  
  bp <- function(x,z){RColorBrewer::brewer.pal(n = 8,name =x)[z]}
  cols <- c(bp('Blues', 8),
            bp('Blues', 5),
            bp('Greys', 3),
            bp('Greys', 6),
            bp('Oranges', 5),
            bp('Oranges', 8))
  if(return_table){
    return(plot_data)
  }
  
  # Remove the ns/nc
  plot_data <- plot_data %>%
    filter(axis != 'NS/NC')
  ggplot(data = plot_data,
         aes(x = axis,
             y = y,
             fill = x)) +
    geom_bar(stat = 'identity',
             position = position_stack(),
             width = 0.9) + 
    # facet_wrap(~ccaa) +
    scale_fill_manual(name = '',
                      values = cols) +
    theme_vilaweb() +
    theme(axis.text.x = element_text(angle = 90, hjust = 1,
                                     vjust = 0.5),
          legend.text = element_text(size = 13)) +
    theme(legend.position = 'right') +
    guides(fill=guide_legend(ncol=1)) +
    geom_hline(yintercept = 50,
               lty = 2,
               alpha = 0.5) +
    labs(x = x,
         y = y,
         title = title,
         subtitle = subtitle,
         caption = caption)
}

left_right_plot <- function(language = 'en',
                            return_table = FALSE){
  if(language == 'en'){
    
    x = 'Ideology (self-assessed)'
    y = 'Percentage'
    title = 'Support for the Spanish Constitution among Catalans'
    subtitle = 'If a referendum were to be held today on the current\nSpanish Constitution approved in 1978, as it is now, how would you vote?'
    caption = 'Data from the Baròmetre d\'Opinió Pública, 3a onada 2018.\nSample size: 1500 residents of Catalonia with Spanish citizenship. Categories codes as follows:\n ideological scale 0-10: 0-1=far left;2-3=left;4-6=center;7-8=right;9-10=far right.\nChart created by Joe Brew | @joethebrew.'
    
    new_axis <- 
      c('Far left',
        'Far left',
        'Left',
        'Left',
        'Center',
        'Center',
        'Center',
        'Right',
        'Right',
        'Far right',
        'Far right')
  } else {
    
    x = 'Ideologia autoubicada (escala 0-10)'
    y = 'Percentage'
    title = 'Suport de la constitució espanyola entre catalans'
    subtitle = 'Si es tornés a celebrar un referèndum per decidir sobre l’actual\nConstitució espanyola aprovada el 1978, tal com és ara, vostè què faria?'
    caption = 'Dades del Baròmetre d\'Opinió Pública, 3a onada 2018.\nMostra: 1500 residents de Catalunya amb ciutadania espayola. Codificació de categories:\nescala ideòlogica 0-10: 0-1=extrema esquerra;2-3=esquerra;4-6=centre;7-8=dreta;9-10=extrema dreta.\nElaboració del gràfic: Joe Brew | @joethebrew.'
    new_axis <- 
      c('Extrema esquerra',
        'Extrema esquerra',
        'Esquerra',
        'Esquerra',
        'Centre',
        'Centre',
        'Centre',
        'Dreta',
        'Dreta',
        'Extrema dreta',
        'Extrema dreta')
  }
  axis_dict <- data_frame(axis = c('Extrema esquerra',
                                   as.character(1:9),
                                   'Extrema dreta'),
                          new_axis) %>%
    mutate(new_axis = gsub(' ', '\n', new_axis))
  plot_data <- 
    vilaweb::ceo %>%
    filter(MES %in% 10:11,
           ANY == 2018) %>%
    dplyr::mutate(axis = P25,
                  constitution = P95) %>%
    filter(!axis %in% c('No ho sap', 'No contesta')) %>%
    left_join(axis_dict) %>%
    dplyr::select(-axis) %>%
    dplyr::rename(axis = new_axis) %>%
    mutate(constitution = as.character(constitution))
  
  if(language == 'en'){
    plot_data <- plot_data %>%
      mutate(constitution = 
               ifelse(constitution == 'Votaria sí', 'Yes',
                      ifelse(constitution == 'Votaria no', 'No',
                             'Not sure/no answer')))
  } else {
    plot_data <- plot_data %>%
      mutate(constitution = ifelse(constitution %in% c('Votaria sí',
                                                       'Votaria no'),
                                   constitution,
                                   'NS/NC/Nul'))
  }
  plot_data <- plot_data %>%
    group_by(axis, constitution)%>%
    summarise(n = sum(PONDERA)) %>%
    ungroup %>%
    group_by(axis) %>%
    mutate(p = n / sum(n) * 100,
           number = sum(n),
           pp = n / sum(n[!constitution %in% c('NS/NC/Nul', 'Not sure/no answer')]) * 100) %>%
    ungroup
  
  if(language == 'en'){
    plot_data$constitution <-
      factor(plot_data$constitution,
             levels = rev(c('Yes',
                            'Not sure/no answer',
                            'No')))
    plot_data$axis <- 
      factor(plot_data$axis,
             levels = c('Far\nleft',
                        'Left',
                        'Center',
                        'Right',
                        'Far\nright'))
  } else {
    plot_data$constitution <-
      factor(plot_data$constitution,
             levels = rev(c('Votaria sí',
                            'NS/NC/Nul',
                            'Votaria no')))
    plot_data$axis <- 
      factor(plot_data$axis,
             levels = c('Extrema\nesquerra',
                        'Esquerra',
                        'Centre',
                        'Dreta',
                        'Extrema\ndreta'))
  }
  
  
  bp <- function(x,z){RColorBrewer::brewer.pal(n = 8,name =x)[z]}
  cols <- rev(c(bp('Blues', 5),
                bp('Greys', 2),
                bp('Oranges', 5)))
  if(return_table){
    return(plot_data)
  }
  ggplot(data = plot_data,
         aes(x = axis,
             y = p,
             fill = constitution)) +
    geom_bar(stat = 'identity',
             # color = 'black',
             # lwd = 0.2,
             position = position_stack()) +
    theme_vilaweb() +
    labs(x = x,
         y = y,
         title = title,
         subtitle = subtitle,
         caption = caption) +
    scale_fill_manual(name = '',
                      values = cols) +
    geom_text(aes(label = round(p, digits = 1)),
              position = position_stack(),
              vjust = 1.4,
              size = 3,
              alpha = 0.6) +
    theme(plot.title = element_text(size = 16),
          plot.subtitle = element_text(size = 11),
          axis.text.x = element_text(size = 15),
          axis.text.y = element_text(size = 13),
          legend.text = element_text(size = 20),
          plot.caption = element_text(size = 10))
}

bp <- function(x = 'Spectral', n = 9){
  RColorBrewer::brewer.pal(n = 9,
                           name = x)[n]
}

referendum_plot <- function(language = 'en',
                            return_table = FALSE){
  plot_data <- vilaweb::ceo %>%
    filter(!is.na(P56K)) %>%
    mutate(P56K = as.character(P56K)) %>%
    mutate(P56K = ifelse(P56K %in% c('No ho sap', 'No contesta'),
                         'NS/NC', P56K)) %>%
    group_by(referendum = P56K) %>%
    summarise(n = sum(PONDERA)) %>%
    ungroup %>%
    mutate(p = n / sum(n) * 100)
  
  right <- 
    data_frame(
      referendum = c("Molt d'acord",
                     "D'acord",
                     "Ni d'acord ni en desacord",
                     "En desacord",
                     "Molt en desacord",
                     "NS/NC"),
      ca = c("Molt\nd'acord",
                     "D'acord",
                     "Ni d'acord\nni en desacord",
                     "En\ndesacord",
                     "Molt en\ndesacord",
                     "NS/NC"),
      en = c("Strongly\nagree",
                          "Agree",
                          "Neither agree\nnor disagree",
                          "Disagree",
                          "Strongly\ndisagree",
                          "No answer/\ndon't know")
    )
  
  plot_data <- left_join(plot_data, right)
  
  x <-''
  y <- 'Percentage'
  if(language == 'en'){
    plot_data$en <-
      factor(plot_data$en,
             levels = c('Strongly\nagree',
                        'Agree',
                        "No answer/\ndon't know",
                        "Neither agree\nnor disagree",
                        'Disagree',
                        'Strongly\ndisagree'))
    plot_data$x <- plot_data$en
    title <- 'Agreement with following phrase:'
    subtitle <- '"Catalonia does not have a right to celebrate a self-determination referendum"'
    caption <- 'Data from Barometer of Public Opinion, 2018, round 2.\nChart by Joe Brew | @joethebrew.'
  } else {
    plot_data$ca <-
      factor(plot_data$ca,
             levels = c("Molt\nd'acord",
                        "D'acord",
                        "Ni d'acord\nni en desacord",
                        'NS/NC',
                        'En\ndesacord',
                        'Molt en\ndesacord'))
    plot_data$x <- plot_data$ca
    title <- "Grau d'acord amb la frase següent"
    subtitle <- '"Catalunya no té el dret de celebrar un referèndum d’autodeterminació"'
    caption <- 'Dades del Baròmetre d\'Opinió Públic, 2018, 2 ronda.\nGràfic de Joe Brew | @joethebrew.'
  }
  if(return_table){
    return(plot_data)
  }
  
  cols <- 
    c(bp('Oranges', c(7,4)),
      bp('Greys', c(4,6)),
      bp('Blues', c(5,7)))
  ggplot(data = plot_data,
         aes(x = x,
             y = p,
             fill = x)) +
    geom_bar(stat = 'identity') +
    theme_vilaweb() +
    labs(x = x,
         y = y,
         title = title,
         subtitle = subtitle,
         caption = caption) +
    scale_fill_manual(name = '',
                      values = cols) +
    theme(legend.position = 'none') +
    geom_text(aes(label = round(p, digits = 1)),
              alpha = 0.7,
              nudge_y = -2,
              color = 'white')
}

simple_plot <- function(language = 'en',
                            return_table = FALSE){
  x <- ''
  if(language == 'en'){
    y = 'Percentage'
    title = 'Support for the Spanish Constitution among Catalans'
    subtitle = 'If a referendum were to be held today on the current\nSpanish Constitution approved in 1978, as it is now, how would you vote?'
    caption = 'Data from the Baròmetre d\'Opinió Pública, 3a onada 2018.\nSample size: 1500 residents of Catalonia with Spanish citizenship.\nChart created by Joe Brew | @joethebrew.'
    
  } else {
    
    y = 'Percentage'
    title = 'Suport de la constitució espanyola entre catalans'
    subtitle = 'Si es tornés a celebrar un referèndum per decidir sobre l’actual\nConstitució espanyola aprovada el 1978, tal com és ara, vostè què faria?'
    caption = 'Dades del Baròmetre d\'Opinió Pública, 3a onada 2018.\nMostra: 1500 residents de Catalunya amb ciutadania espayola.\nElaboració del gràfic: Joe Brew | @joethebrew.'
  }

  plot_data <- 
    vilaweb::ceo %>%
    filter(MES %in% 10:11,
           ANY == 2018) %>%
    dplyr::mutate(constitution = P95) %>%
    mutate(constitution = as.character(constitution))
  
  if(language == 'en'){
    plot_data <- plot_data %>%
      mutate(constitution = 
               ifelse(constitution == 'Votaria sí', 'Yes',
                      ifelse(constitution == 'Votaria no', 'No',
                             'Not sure/no answer')))
  } else {
    plot_data <- plot_data %>%
      mutate(constitution = ifelse(constitution %in% c('Votaria sí',
                                                       'Votaria no'),
                                   constitution,
                                   'NS/NC/Nul'))
  }
  plot_data <- plot_data %>%
    group_by(constitution)%>%
    summarise(n = sum(PONDERA)) %>%
    ungroup %>%
    mutate(p = n / sum(n) * 100,
           number = sum(n),
           pp = n / sum(n[!constitution %in% c('NS/NC/Nul', 'Not sure/no answer')]) * 100) %>%
    ungroup
  
  if(language == 'en'){
    plot_data$constitution <-
      factor(plot_data$constitution,
             levels = rev(c('Yes',
                            'Not sure/no answer',
                            'No')))
  } else {
    plot_data$constitution <-
      factor(plot_data$constitution,
             levels = rev(c('Votaria sí',
                            'NS/NC/Nul',
                            'Votaria no')))
  }
  
  
  bp <- function(x,z){RColorBrewer::brewer.pal(n = 8,name =x)[z]}
  cols <- rev(c(bp('Blues', 5),
                bp('Greys', 6),
                bp('Oranges', 5)))
  if(return_table){
    return(plot_data)
  }
  ggplot(data = plot_data,
         aes(x = constitution,
             y = p,
             fill = constitution)) +
    geom_bar(stat = 'identity',
             # color = 'black',
             # lwd = 0.2,
             position = position_stack()) +
    theme_vilaweb() +
    labs(x = x,
         y = y,
         title = title,
         subtitle = subtitle,
         caption = caption) +
    scale_fill_manual(name = '',
                      values = cols) +
    geom_text(aes(label = round(p, digits = 1)),
              position = position_stack(),
              vjust = 2,
              # size = 3,
              color = 'white',
              alpha = 0.7) +
    theme(plot.title = element_text(size = 16),
          plot.subtitle = element_text(size = 11),
          axis.text.x = element_text(size = 15),
          axis.text.y = element_text(size = 13),
          legend.text = element_text(size = 20),
          plot.caption = element_text(size = 10),
          legend.position = 'none')
}
# simple_plot()

self_determination_plot <- function(language = 'en',
                                    geo = 'cat',
                                    return_table = FALSE){
  cis_data  <- cis_list$`2018-10`
  if(geo == 'esp'){
    plot_data <- cis_data %>% filter(CCAA != 'Cataluña')  
    if(language == 'en'){subtitle <- 'Spain, not counting Catalonia'} else {subtitle <-'Espanya, sense incloure Catalunya'}
  }
  if(geo == 'cat'){
    plot_data <- cis_data %>% filter(CCAA == 'Cataluña')
    if(language == 'en'){subtitle <- 'Catalonia'} else {subtitle <-'Catalunya'}
  }
  if(geo == 'all'){
    if(language == 'en'){subtitle <- 'Spain, including Catalonia'} else {subtitle <- 'Espanya, incloent Catalonia'}
    plot_data <- cis_data 
  }
  
  right <- 
    data_frame(P26 = 
                 c('Un Estado con un único Gobierno central sin autonomías',
                   'Un Estado en el que las comunidades autónomas tengan menor autonomía que en la actualidad',
                   'Un Estado con comunidades autónomas como en la actualidad',
                   'Un Estado en el que las comunidades autónomas tengan mayor autonomía que en la actualidad',
                   'Un Estado en el que se reconociese a las comunidades autónomas la posibilidad de convertirse en Estados independientes',
                   'N.S.',
                   'N.C.'),
               en = c('Central State/\nNo autonomy',
                      'Less autonomy',
                      'Status Quo',
                      'More autonomy',
                      'Possibility of independence',
                      'No answer/\nnot sure',
                      'No answer/\nnot sure'),
               ca = c('Estat central/\nSense autonomia',
                      'Menys autonomia',
                      'Statu Quo',
                      'Mes autonomia',
                      'Possibilitat de\n independència',
                      'NS/NC',
                      'NS/NC'))
  
  plot_data <- plot_data %>%
    left_join(right)
  if(language == 'en'){
    plot_data$x <- plot_data$en
    title <- 'Preferences for territorial organization'
    caption <- 'Data from CIS survey, October 2018.\nJoe Brew | @joethebrew.'
    x <- 'Percentage'
    y <- ''
  } else {
    plot_data$x <- plot_data$ca    
    title <- 'Preferències sobre organització territorial'
    caption <- 'Dades del CIS, Octubre 2018.\nJoe Brew | @joethebrew.'
    x <- 'Percentage'
    y <- ''
  }
  plot_data <- plot_data %>%
    group_by(x) %>%
    tally %>%
    ungroup %>%
    mutate(p = n / sum(n) * 100)
  if(language == 'en'){
    plot_data$x <- factor(plot_data$x, levels =
                            c('Central State/\nNo autonomy',
                              'Less autonomy',
                              'Status Quo',
                              'No answer/\nnot sure',
                              'More autonomy',
                              'Possibility of independence'))
  } else {
    plot_data$x <- factor(plot_data$x, levels =
                            c('Estat central/\nSense autonomia',
                              'Menys autonomia',
                              'Statu Quo',
                              'NS/NC',
                              'Mes autonomia',
                              'Possibilitat de\n independència'))
  }
  
  cols <- c(bp('Blues', 8),
            bp('Blues', 5),
            bp('Greys', 5),
            bp('Greys', 7),
            bp('Oranges', 5),
            bp('Oranges', 8))
  
  plot_data <- plot_data %>%
    arrange(x)
  
  if(return_table){
    return(plot_data)
  }

  ggplot(data = plot_data,
         aes(x = x,
             y = p,
             fill = x)) +
    geom_bar(stat = 'identity') +
    geom_text(aes(label = round(p, digits = 1)),
              alpha = 0.7,
              color = 'white',
              nudge_y = -2) +
    theme_vilaweb()+
    scale_fill_manual(name = '',
                      values = cols) +
    theme(legend.position = 'none') +
    coord_flip() +
    labs(x = x,
         y = y,
         title = title,
         subtitle = subtitle,
         caption= caption)
  
}


make_franco <- function(){
  
  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",
                      "Cs",
                      "Reagrupament.cat",
                      "SI",
                      "PxC",
                      "CUP",
                      "UPyD",
                      "Podem",
                      "Podem",
                      "PDCat/CiU/CDC/Junts",
                      "PDCat/CiU/CDC/Junts",
                      "Podem",
                      "Democràcia i Llibertat",
                      "Podem",
                      "PACMA",
                      "PDCat/CiU/CDC/Junts",
                      "PDCat/CiU/CDC/Junts",
                      "Podem",
                      "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') %>%
  
  # Get date
  pd <- pd %>%
    arrange(ANY, MES) %>%
    group_by(BOP_NUM) %>%
    mutate(bop_date = as.Date(paste0(dplyr::first(ANY),
                                     '-',
                                     dplyr::first(MES),
                                     '-01'))) 
  
  # # Get valoracio del rei 2018 onada 1 p21j
  # pd <- pd %>%
  #   mutate(valoracio_rei = P21J)
  # 
  # # Get val columns
  val_columns <- names(pd)[grepl('val_', names(pd))]
  
  # Franquism question: P102
  pd$franquisme <- pd$P102
  pd$constitucio <- pd$P95
  con_dict <- data_frame(
    constitucio = c('Votaria sí',
                    'Votaria no',
                    'Votaria en blanc',
                    'Votaria nul',
                    'No votaria',
                    'No ho sap',
                    'Contesta'),
    constitucio2 = c('Votaria sí',
                     'Votaria no',
                     'Votaria nul\nen blanc\nNo votaria\nNS/NC',
                     'Votaria nul\nen blanc\nNo votaria\nNS/NC',
                     'Votaria nul\nen blanc\nNo votaria\nNS/NC',
                     'Votaria nul\nen blanc\nNo votaria\nNS/NC',
                     'Votaria nul\nen blanc\nNo votaria\nNS/NC')
  )
  pd <- pd %>%
    left_join(con_dict) %>%
    dplyr::select(-constitucio) %>%
    dplyr::rename(constitucio = constitucio2)
  
  # Subset to only include columns of interest
  pd <- pd[,c('axis', 
              'bop_date',
              'partit',
              'indy',
              'franquisme',
              'constitucio',
              'PONDERA')]
  pd <- pd %>%
    filter(!is.na(franquisme)) %>%
    mutate(franquisme = as.character(franquisme)) %>%
    mutate(franquisme = ifelse(franquisme %in% c('No ho sap', 'No contesta'), 'NS/NC', franquisme)) %>%
    mutate(franquisme = ifelse(franquisme == 'Va tenir coses positives i negatives', 'Va tenir\ncoses positives\ni negatives', franquisme))
  pd$franquisme <- factor(pd$franquisme,
                          levels = c('Negatiu',
                                     'Va tenir\ncoses positives\ni negatives', 
                                     'Positiu',
                                     'NS/NC'))
  
  # # Gather
  # pd <- pd %>%
  #   gather(key, value, val_Albiol:val_Albiach)
  plot_data <- pd %>%
    group_by(partit, franquisme) %>%
    summarise(n = sum(PONDERA),
              mostra = n()) %>%
    group_by(partit) %>%
    mutate(p = n / sum(n) * 100) %>%
    ungroup %>%
    filter(!partit %in% c('ICV-EUiA', 'PACMA',
                          'Altre/Cap/NS/NC'))
  
  plot_data <- plot_data %>%
    arrange(franquisme,p)
  plot_data$partit <- gsub('/', '/\n', plot_data$partit)
  plot_data$partit <- factor(plot_data$partit,
                             levels = unique(plot_data$partit))
  # plot_data$franquisme <- factor(plot_data$franquisme)
  # plot_data$franquisme <- factor(plot_data$franquisme,
  #                                levels = rev(levels(plot_data$franquisme)))
  
  cols <- RColorBrewer::brewer.pal(n = 8, 'Spectral' )
  cols <- cols[1:4]
  cols <- rev(cols)
  
  # cols <- rev(cols)
  cols[4] <- grey(0.6)
  cols[1] <- as.character(vilaweb::colors_vilaweb()[5])
  
  plot_data$label <- ifelse(plot_data$franquisme == 'NS/NC',
                            '',
                            paste0(round(plot_data$p, 1), '%'))
  
  ggplot(data = plot_data,
         aes(x = partit,
             y = p,
             fill = franquisme)) +
    geom_bar(stat = 'identity',
             position = position_stack()) +
    scale_fill_manual(name = '',
                      values = cols) +
    geom_text(aes(label = label),
              # nudge_y = 4,
              size = 3,
              position = position_stack(),
              color = 'white',
              alpha = 0.8,
              vjust = 1) +
    labs(y = 'Percentage',
         x = '',
         title = 'Valoració del franquisme per partit',
         subtitle = "Pregunta: 'Per a vostè, l’etapa del franquisme ha suposat a la història de Catalunya, \nen conjunt, un període positiu, negatiu o va tenir coses positives i negatives?'",
         caption = paste0(sum(plot_data$mostra),
                          ' residents de Catalunya amb ciutadania espanyola. Dades del BOP/CEO 3a onada 2018.\nPreguntes 102 (franquisme) i 24 (partit). Gràfic de Joe Brew | @joethebrew. |  www.vilaweb.cat.')) +
    scale_fill_manual(name = 'El franquisme\nva ser...',
                      values = cols) +
    theme_vilaweb() +
    theme(legend.position = 'right') +
    # theme(legend.position = 'none') +
    theme(axis.text.x = element_text(size = 10),
          axis.text.y = element_text(size = 14),
          plot.title = element_text(hjust = 0.5, size = 24),        plot.subtitle = element_text(size = 10),
          
          strip.text = element_text(size = 21,
                                    color = grey(0.1)),
          plot.caption = element_text(hjust = 0, size = 7))
}

make_plot <- function(var = 'P56J',
                      only_2018 = FALSE,
                      show_labels = TRUE,
                      ceo = vilaweb::ceo){
  
  df <- ceo
  if(only_2018){
    df <- df %>%
      filter(ANY == 2018)
  }
  df <- df %>%
    mutate(axis = P25,
           indy = P31) %>%
    mutate(indy = as.character(indy)) %>%
    mutate(indy = ifelse(indy %in% c('No ho sap',
                                     'No contesta'),
                         'NS/NC',
                         indy)) %>%
    group_by(axis, indy) %>%
    tally %>%
    ungroup %>%
    group_by(axis) %>%
    mutate(p = n / sum(n) * 100) %>%
    # filter(!axis %in% c('No ho sap', 'No contesta')) %>%
    # filter(!is.na(indy)) %>%
    mutate(`Muestra` = n)
  
  df$axis <- factor(df$axis,
                    levels = levels(df$axis),
                    labels = gsub(' ', '\n', levels(df$axis)))
  
  zz <- theme(axis.text.x = element_text(size = 10),
              plot.caption  = element_text(hjust = 0))
  
  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",
                      "Cs",
                      "Reagrupament.cat",
                      "SI",
                      "PxC",
                      "CUP",
                      "UPyD",
                      "Podem",
                      "Podem",
                      "PDCat/CiU/CDC/Junts",
                      "PDCat/CiU/CDC/Junts",
                      "Podem",
                      "Democràcia i Llibertat",
                      "Podem",
                      "PACMA",
                      "PDCat/CiU/CDC/Junts",
                      "PDCat/CiU/CDC/Junts",
                      "Podem",
                      "Altre/Cap/NS/NC",
                      "Altre/Cap/NS/NC",
                      "Altre/Cap/NS/NC",
                      "Altre/Cap/NS/NC")) %>%
    mutate(partit = ifelse(partit == "PDCat/CiU/CDC/Junts",
                           "PDCat/CiU/\nCDC/Junts",
                           partit)) %>%
    mutate(partit = ifelse(partit == 'Podem',
                           'Podem',
                           partit))
  
  
  pd <- vilaweb::ceo
  pd$var <- as.character(pd[,var] %>% unlist)
  
  pd <- pd %>%
    left_join(party_dict) %>%
    filter(partit %in% c("Cs", "CUP", "ERC",
                         "Podem", "PPC",'PSC', 'PDCat/CiU/\nCDC/Junts')) %>%
    mutate(indy = partit) %>%
    mutate(var = as.character(var)) %>%
    mutate(var = ifelse(var %in% c('No ho sap',
                                   'No contesta'),
                        'NS/NC',
                        var)) %>%
    filter(!is.na(var)) %>%
    filter(var != 'NS/NC') %>%
    mutate(var = 
             ifelse(var %in% c("Molt d'acord",
                               "D'acord"),
                    "D'acord o\nmolt d'acord",
                    ifelse(var %in% c("En desacord",
                                      "Molt en desacord"),
                           "En desacord o\nmolt en desacord",
                           "Ni d'acord ni\nen desacord"))) %>%
    group_by(var, indy) %>%
    summarise(n = sum(PONDERA),
              people = n()) %>%
    ungroup %>%
    group_by(indy) %>%
    mutate(p = n / sum(n) * 100) %>%
    mutate(`Muestra` = n)
  
  pd <- pd %>% arrange(var, p)
  pd$indy <- factor(pd$indy, levels = unique(pd$indy))
  
  phrase_dict <- 
    tibble(var = paste0('P56',
                        c(LETTERS[1:11])),
           catalan = c("'Com menys intervingui el govern en l’economia, millor serà pel país'",
                       "'Cal abaixar els impostos, encara que això impliqui\nreduir serveis i prestacions públiques'",
                       "'El govern hauria de prendre mesures per a reduir les\ndiferències en els nivells d’ingressos'",
                       "'Les parelles de gais i lesbianes han de poder adoptar\nfills en les mateixes condicions que les parelles heterosexuals'",
                       "'L’escola ha d’ensenyar als nens a obeir l’autoritat'",
                       "'La religió no hauria de tenir cap influència en la política'",
                       "'En qualsevol circumstància, la llei sempre ha de ser obeïda'",
                       "'Algú amb plenes facultats hauria de poder decidir quan vol morir'",
                       "'Amb tanta immigració, un ja no se sent com a casa'",
                       "'El creixement econòmic ha de tenir prioritat sobre la protecció del medi ambient'",
                       "'Catalunya no té el dret de celebrar un referèndum d’autodeterminació'"),
           english = c("'The less the government interferes in the economy, the better off the country will be'",
                       "'Taxes must be lowered, even though it may\nmean reducing public services'",
                       "'The government should take measures to\nreduce differenes in income'",
                       "'Gay and lesbian couples should be able to adopt children\nunder the same conditions as heterosexual couples'",
                       "'School should teach children to obey authority'",
                       "'Religion should have no influence on politics'",
                       "'The law should always be obeyed in any circumstance'",
                       "'Someone with full abilities should be allowed to decide when (s)he wants to die'",
                       "'With so much immigration, one no longer feels at home'",
                       "'Economic growth should have priority over protection of the environment'",
                       "'Catalonia does not have a right to hold a self-determination referendum'"))
  
  the_phrase <- phrase_dict$catalan[phrase_dict$var == var]
  line_breaker <- function(x){
    x <- stri_wrap(x, width = 50)
    x <- paste0(x, collapse = '\n')
    return(x)
  }
  the_phrase <- line_breaker(the_phrase)  
  pd$var<- factor(pd$var,
                  levels = c("D'acord o\nmolt d'acord",
                             "Ni d'acord ni\nen desacord",
                             "En desacord o\nmolt en desacord"),
                  labels = c("D'acord",
                             "Ni d'acord ni\nen desacord",
                             "En desacord"))
  n_cols <- length(unique(pd$var))
  # cols <- databrew::make_colors(n = n_cols, categorical = FALSE)
  # cols <- as.character(vilaweb::colors_vilaweb()[c(1,5,3)])
  cols <- c(
    RColorBrewer::brewer.pal(n = 9, name = 'Oranges')[6],
    'darkgrey',
    RColorBrewer::brewer.pal(n = 9, name = 'Blues')[7]
  )
  cols <- rev(cols)
  
  if(only_2018){
    the_caption <- paste0('Mostra: ', point_replace(commafy(sum(pd$people))),
                          ' residents de Catalunya amb ciutadania espanyola.\nEnquestes BOP del CEO, 2018. Preguntes ',
                          var,
                          ' i P31.\nJoe Brew | @joethebrew. | www.vilaweb.cat')
  } else {
    the_caption <- paste0('Mostra: ', point_replace(commafy(sum(pd$people))),
                          ' residents de Catalunya amb ciutadania espanyola.\nCombinació enquestes BOP del CEO, 2015 i 2018. Preguntes ',
                          var,
                          ' i P31.\nPercentatges calculats amb ponderació. Joe Brew | @joethebrew. | www.vilaweb.cat')
  }
  
  g <- ggplot(data = pd,
              aes(x = indy,
                  y = p)) +
    geom_bar(stat = 'identity',
             position = position_stack(vjust = 0.5, reverse = T),
             # color = 'black',
             alpha = 0.9,
             aes(fill = var,
                 group = var)) +
    theme_vilaweb() +
    scale_fill_manual(name = '',
                      values = cols) +
    labs(x = '',
         y = 'Percentage',
         subtitle = "Grau d'acord amb l'afirmació:",
         title =  the_phrase,
         caption = the_caption) +
    theme(legend.position = 'right') +
    zz +
    guides(fill = guide_legend(reverse = T))
  if(show_labels){
    g <- g +
      geom_text(aes(label = round(p, digits = 2),
                    group = var),
                position = position_stack(vjust = 0.5,
                                          reverse = T),
                alpha = 0.7,
                color = 'white',
                size = 3) 
  }
  return(g)
}
library(tidyverse)
library(vilaweb)
party_position <- function(snapshot = FALSE){
  
  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",
                      "Cs",
                      "Reagrupament.cat",
                      "SI",
                      "PxC",
                      "CUP",
                      "UPyD",
                      "Podem",
                      "Podem",
                      "PDCat/CiU/CDC/Junts",
                      "PDCat/CiU/CDC/Junts",
                      "Podem",
                      "Democràcia i Llibertat",
                      "Podem",
                      "PACMA",
                      "PDCat/CiU/CDC/Junts",
                      "PDCat/CiU/CDC/Junts",
                      "Podem",
                      "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'))
  # P25
  
  if(snapshot){
    agg <- pd %>%
      filter(ANY == 2018)
      nn <- nrow(agg[!is.na(agg$axis),])
    
    agg <- agg %>%
      group_by(partit) %>%
      summarise(avg = weighted.mean(x = axis, w = PONDERA, na.rm = TRUE)) %>%
      ungroup
    
    cols <- length(unique(agg$partit))
    cols <- c('darkorange', 'yellow', 'black',
              grey(0.6), 'darkgreen', 'darkred',
              'purple', 'blue', 'red')
    cols <- cols[order(agg$avg)]
  agg <- agg %>% arrange(avg)
  agg$partit <- gsub('PDCat/CiU/CDC/Junts', 'PDCat/Junts', agg$partit)
  agg$partit <- factor(agg$partit, levels = agg$partit)  
  
  g <- ggplot(data = agg,
              aes(x = partit,
                  y = avg)) +
    geom_bar(stat = 'identity',
             aes(fill = partit)) +
    scale_fill_manual(name = '', values = cols) +
    theme_vilaweb() +
    theme(legend.position = 'none') +
    coord_flip() +
    labs(x = '',
         y = 'Resposta mitjana: escala esquerra (0) - dreta (10)') +
    geom_text(aes(label = round(avg, digits = 1)),
              nudge_y = -0.3,
              color = 'white')
  
  } else {
    nn <- nrow(pd[!is.na(pd$axis),])
    agg <- pd %>%
      group_by(partit, year) %>%
      summarise(avg = weighted.mean(x = axis, w = PONDERA, na.rm = TRUE)) %>%
      ungroup
    
    cols <- length(unique(agg$partit))
    cols <- c('darkorange', 'yellow', 'black',
              grey(0.6), 'darkgreen', 'darkred',
              'purple', 'blue', 'red')
    # cols <- rainbow(cols)
    g <- ggplot(data = agg,
           aes(x = year,
               y = avg)) +
      geom_bar(stat = 'identity',
               alpha = 0.6,
               aes(fill = partit)) +
      theme_vilaweb() +
      facet_wrap(~partit) +
      theme(axis.text.x = element_text(angle = 90,
                                       vjust = 0.5,
                                       hjust = 1)) +
      geom_text(aes(label = round(avg, digits = 1)),
                nudge_y = 1,
                size = 2) +
      coord_flip() +
      scale_fill_manual(name = '',values = cols) +
      theme(legend.position = 'none') +
      geom_point() +
      geom_line() +
      labs(x = 'Any',
           y = 'Escala esquerra (0) / dreta (10)') +
      ylim(0, 10) +
      scale_y_continuous(breaks = c(0, 2,4,6)) +
      geom_hline(yintercept = 5, lty = 2, alpha = 0.3)
  }
  
  g <- g +
    labs(subtitle = "Pregunta: 'Em pot dir on s’ubicaria vostè en una escala de 0 a 10 on\n0 significa extrema esquerra i 10 extrema dreta?'",
  caption = paste0(point_replace(commafy(nn)),
                   ' residents de Catalunya amb ciutadania espanyola. Dades del BOP/CEO\nPregunta P25 (ideologia) i P24 (partit). Gràfic de Joe Brew | @joethebrew. |  www.vilaweb.cat.'),
  title = 'Autoubicació ideològica per partit') +
    theme(plot.subtitle = element_text(hjust = 0))
  return(g)
}

make_val <- function(who = 'Albiach',
                     ceo = vilaweb::ceo,
                     only_2018 = FALSE){
  
  df <- ceo
  if(only_2018){
    df <- df %>%
      filter(ANY == 2018)
  }
  df <- df %>%
    mutate(axis = P25,
           indy = P31) %>%
    mutate(indy = as.character(indy)) %>%
    mutate(indy = ifelse(indy %in% c('No ho sap',
                                     'No contesta'),
                         'NS/NC',
                         indy)) %>%
    group_by(axis, indy) %>%
    tally %>%
    ungroup %>%
    group_by(axis) %>%
    mutate(p = n / sum(n) * 100) %>%
    # filter(!axis %in% c('No ho sap', 'No contesta')) %>%
    # filter(!is.na(indy)) %>%
    mutate(`Muestra` = n)
  
  df$axis <- factor(df$axis,
                    levels = levels(df$axis),
                    labels = gsub(' ', '\n', levels(df$axis)))
  
  zz <- theme(axis.text.x = element_text(size = 10),
              plot.caption  = element_text(hjust = 0))
  
  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",
                      "Cs",
                      "Reagrupament.cat",
                      "SI",
                      "PxC",
                      "CUP",
                      "UPyD",
                      "Podem",
                      "Podem",
                      "PDCat/CiU/CDC/Junts",
                      "PDCat/CiU/CDC/Junts",
                      "Podem",
                      "Democràcia i Llibertat",
                      "Podem",
                      "PACMA",
                      "PDCat/CiU/CDC/Junts",
                      "PDCat/CiU/CDC/Junts",
                      "Podem",
                      "Altre/Cap/NS/NC",
                      "Altre/Cap/NS/NC",
                      "Altre/Cap/NS/NC",
                      "Altre/Cap/NS/NC")) %>%
    mutate(partit = ifelse(partit == "PDCat/CiU/CDC/Junts",
                           "PDCat/CiU/\nCDC/Junts",
                           partit)) %>%
    mutate(partit = ifelse(partit == 'Podem',
                           'Podem',
                           partit))
  
  
  pd <- vilaweb::ceo
  var <- paste0('val_', who)
  pd$var <- as.character(pd[,var] %>% unlist)
  pd$var <- as.character(pd$var)
  pd$var <- ifelse(pd$var == 'Excel·lent', '10',
                   ifelse(pd$var == 'Molt deficient', 0,
                          pd$var))
  pd$var <- as.numeric(pd$var)
  pd <- pd %>%
    left_join(party_dict) %>%
    filter(partit %in% c("Cs", "CUP", "ERC",
                         "Podem", "PPC",'PSC', 'PDCat/CiU/\nCDC/Junts')) %>%
    mutate(indy = partit) %>%
    filter(!is.na(var)) %>%
    group_by(indy) %>%
    summarise(avg = mean(var, na.rm = TRUE),
              n = sum(PONDERA),
              people = n()) %>%
    ungroup %>%
    mutate(p = n / sum(n) * 100) %>%
    mutate(`Muestra` = n) %>%
    arrange(avg)
  pd$indy <- factor(pd$indy, levels = pd$indy)
  
  ggplot(data = pd,
         aes(x = indy,
             y = avg)) +
    geom_bar(stat = 'identity',
             fill = 'darkorange',
             alpha = 0.7) +
    geom_text(aes(label = round(avg, digits = 1)),
              color = 'white',
              nudge_y = -0.5) +
    labs(title = ifelse(who == 'Albiach',
                        'Valoració d\'Albiach',
                        paste0('Valoració de ', who)),
         caption = paste0(point_replace(commafy(sum(pd$people))),
                          ' residents de Catalunya amb ciutadania espanyola. Dades del BOP/CEO.\nGràfic de Joe Brew | @joethebrew. |  www.vilaweb.cat.', ifelse(who == 'Albiach',
                                                                                                                                                                '\nExclou el PPC perquè cap dels enquestats sabia qui era Albiach, i per tant no podian valorar-la.', '')),
                          x = 'Votants d\'aquest partit',
                          y = 'Valoració (0-10)') +
    theme_vilaweb() +
    theme(plot.subtitle = element_text(hjust = 0))
}