doing_something_with_download.R

library(readr)
library(dplyr)
library(tidyr)
library(ggplot2)
library(ggthemes)
library(lubridate)

# because we have a lot of csv files of mentions, and only want a few columns
# from each, we will us a custom read function for each csv
custom_read <- function(x){
# where x is the path to the csv file, 
# which includes the username as the second last bit (the folder the csv file is in)
  username <- rev(unlist(strsplit(x, split="/")))[2]
  xcontent <- read_csv(x, col_types = 
                         cols_only(created_at = col_datetime(format = ""),
                                   status_id = col_character(),
                                   user_id = col_character()))
  if(nrow(xcontent) > 0){
    xcontent$sent_to <- username
    return(xcontent)
  } else {return(NULL)}
}
# assumes for this script that the working directory has a folder called
# twitter_accounts containing a folders names by twitter screennames,
# contain csv files with download date and the text MN for mentions of
# the account name

readthese <- list.files(path="twitter_accounts",
                        pattern="MN\\.csv",
                        full.names=TRUE,
                        recursive = TRUE)
mntions <- bind_rows(lapply(readthese,custom_read))

weeker <- mntions %>%
  select(user_id, created_at, status_id) %>% 
  distinct() %>% # removing overlap between captures 
  filter(created_at > ISOdatetime(2020,1,1,0,0,0)) %>% #2020 only
  arrange(user_id, created_at) %>%
  group_by(user_id) %>% #
  slice(1) %>%          # gets first contact in 2020
  ungroup() %>%         #
  mutate(week_of = floor_date(created_at, unit="week", week_start = 1),
         week_no = as.numeric(factor(week_of))) # for weeks from Mon
ggplot(weeker, aes(x=week_of)) + geom_bar() +
  xlab("Week (starting) from Monday") +
  theme_tufte() + ggtitle("Total new accounts mentioning
1 or more NZ Plotical Twitter accounts")