affordability.R

library(tidyverse)
library(janitor)
library(zoo)
library(scales)


clean_and_avg = function(series){
  dat = read_csv(sprintf('%s.csv',series))
  
  dat = janitor::clean_names(dat)
  
  dat$date = as.Date(as.POSIXct(dat$unix_timestamp/1000, origin = "1970-01-01", tz = "UTC")) 
  
  
  dat = dat %>%
    group_by(party,date) %>%
    summarise(n = n()) %>%
    group_by(party) %>%
    arrange(party, date) %>%
    mutate(ntot = cumsum(n)) %>%
    group_by(party) %>%
    mutate(daily_avg = rollapply(data = n, width = 30, partial = T, fill = NA, FUN = mean, na.rm = T, align = 'right')) %>%
    ungroup() %>%
    mutate(index = rescale(daily_avg, to = c(0, 100))) %>%
    rename({{series}} := daily_avg) %>%
    select(all_of(c('party','date',series))) 
  
  return(dat)
}


dat = 
  expand_grid(party = c("Democrat","Republican"),
         date = seq.Date(ymd('2010-01-01'),ymd('2025-11-25'))) %>%
  left_join(
    clean_and_avg('inflation'),
    by=c('party','date')
    ) %>%
  left_join(
    clean_and_avg('affordability'),
    by=c('party','date')
  ) %>%
  left_join(
    clean_and_avg('prices'),
    by=c('party','date')
  ) 

dat = dat %>%
  mutate_at(c('inflation','affordability','prices'),
            na.approx, na.rm = F)




dat %>%
  filter(date >= ymd('2024-01-01')) %>%
  gather(word, avg, 3:5) %>%
  filter(word == 'affordability') %>%
  ungroup() %>%
  mutate(index = rescale(avg, to = c(0, 100))) %>%
  ggplot(., aes(x = date, y = index, col = party)) + 
  # annotation one
  annotate(geom='text',x = ymd('2025-06-20'),y=80,
           label = 'Zohran Mamdani wins 2025\nDemocratic mayoral primary\nin New York City',
           hjust = 1) +
  geom_vline(xintercept = ymd('2025-06-24'),col='black') +
  # annotation 2
  annotate(geom='text',x = ymd('2025-11-01'),y=90,
           label = '2025\nstatewide\nand NYC\ngeneral\nelections',
           hjust = 1) +
  geom_vline(xintercept = ymd('2025-11-04'),col='black') +
  # annotation 3
  annotate(geom='text',x = ymd('2024-11-01'),y=90,
           label = 'Trump wins\n2024 election',
           hjust = 1) +
  geom_vline(xintercept = ymd('2024-11-05'),col='black') +
  # annotation 4
  annotate(geom='text',x = ymd('2024-07-15'),y=60,
           label = 'Joe Biden drops out\nof 2024 presidential election',
           hjust = 1) +
  geom_vline(xintercept = ymd('2024-07-21'),col='black') +
  # rest of chart
  geom_line() + 
  scale_color_manual(values = c('Democrat' = 'blue','Republican'='red')) +
  coord_cartesian(ylim=c(0,100)) +
  scale_x_date(date_labels = '%b %Y') +
  labs(x = 'Date',
       y = 'Mentions index\n(max monthly avg = 100)',
       subtitle = 'Index of monthly average mentions of the word "affordability" in congressional emails,\nby political party. 100 = max monthly value.',
       title='Republicans are playing catch-up on "affordability"',
       col = 'Party',
       caption = 'Data: DCInbox.com\nSource: Strength In Numbers/gelliottmorris.com') + 
  theme_minimal() +
  theme(legend.position = 'top',legend.justification = 'left',
        plot.caption = element_text(hjust=0),
        panel.grid.minor.x = element_blank()) 


dat %>%
  filter(date >= ymd('2025-06-01')) %>%
  gather(word, avg, 3:5) %>%
  filter(word == 'affordability') %>%
  ungroup() %>%
  mutate(index = rescale(avg, to = c(0, 100))) %>%
  ggplot(., aes(x = date, y = index, col = party)) + 
  # annotation one
  annotate(geom='text',x = ymd('2025-06-25'),y=60,
           label = 'Zohran Mamdani wins 2025\nDemocratic mayoral primary\nin New York City',
           hjust = 0) +
  geom_vline(xintercept = ymd('2025-06-24'),col='black') +
  # annotation 2
  annotate(geom='text',x = ymd('2025-11-01'),y=90,
           label = '2025\nstatewide\nand NYC\ngeneral\nelections',
           hjust = 1) +
  geom_vline(xintercept = ymd('2025-11-04'),col='black') +
  # rest of chart
  geom_line() + 
  scale_color_manual(values = c('Democrat' = 'blue','Republican'='red')) +
  coord_cartesian(ylim=c(0,100)) +
  scale_x_date(date_breaks = 'month', date_labels = '%b') +
  labs(x = '2025',
       y = 'Mentions index\n(max monthly avg = 100)',
       subtitle = 'Index of monthly average mentions of the word "affordability" in congressional emails,\nby political party. 100 = max monthly value.',
       title='Republicans are playing catch-up on "affordability"',
       col = 'Party',
       caption = 'Data: DCInbox.com\nSource: Strength In Numbers/gelliottmorris.com') + 
  theme_minimal() +
  theme(legend.position = 'top',legend.justification = 'left',
        plot.caption = element_text(hjust=0),
        panel.grid.minor.x = element_blank()) 




dat %>%
  filter(date >= ymd('2024-01-01')) %>%
  gather(word, avg, 3:5) %>%
  #filter(word != 'prices') %>%
  ungroup() %>%
  mutate(index = rescale(avg, to = c(0, 100))) %>%
  ggplot(., aes(x = date, y = index, col = word)) + 
  geom_line() + 
  facet_wrap(~party,ncol=1) + 
  theme_minimal() +
  scale_color_manual(values = c('affordability' = 'green','inflation'='gray40','prices'='pink')) +
  coord_cartesian(ylim=c(0,100)) +
  scale_x_date(date_breaks = '6 month', date_labels = '%b %Y') +
  labs(x = 'Date',
       y = 'Mentions index\n(max monthly avg = 100)',
       subtitle = 'Index of monthly average mentions of words related to affordability and prices\nin congressional emails, by word political party. 100 = max monthly value.',
       title='"Affordability" is the new "inflation"',
       col = 'Word',
       caption = 'Data: DCInbox.com\nSource: Strength In Numbers/gelliottmorris.com') + 
  theme_minimal() +
  theme(legend.position = 'top',legend.justification = 'left',
        plot.caption = element_text(hjust=0),
        panel.grid.minor.x = element_blank())