calculate cabinet duration with ParlGov data in R -- two approaches

parlgov-cabinet-duration.R

# get and read ParlGov cabinet data
tbl <- 'view_cabinet.csv'

if( ! tbl %in% list.files()) {
  download.file(paste0('http://www.parlgov.org/static/data/development-cp1252/', tbl),  tbl)
}

## Base R approach

cab <- read.csv(tbl, fileEncoding = 'cp1252', as.is = TRUE)

# calculate duration of cabinet through self-merge on cabinet IDs
cab_dur <- cab[ ! duplicated(cab$cabinet_id), c('cabinet_id', 'previous_cabinet_id', 'start_date')]
cab_dur$start_date <- as.Date(cab_dur$start_date)
cab_dur <- merge(cab_dur, cab_dur, by.x = 'cabinet_id', by.y = 'previous_cabinet_id')
cab_dur$duration <- cab_dur$start_date.y - cab_dur$start_date.x

cab <- merge(cab, cab_dur[ , c('cabinet_id', 'duration')], by = 'cabinet_id')


## dplyr approach -- way nicer

library(dplyr)

cab <- read.csv(tbl, fileEncoding = 'cp1252', as.is = TRUE)

cab_dur <- cab %>%
  mutate(start_date = as.Date(start_date)) %>% 
  select(cabinet_id, previous_cabinet_id, start_date) %>%
  distinct

cab_dur <- cab_dur %>%
  inner_join(cab_dur, by = c('cabinet_id' = 'previous_cabinet_id')) %>% 
  mutate(duration = start_date.y - start_date.x) %>% 
  select(cabinet_id, duration)

cab <- cab %>% left_join(cab_dur)