Analyze email traffic between two people and animate it

demo_gmailr.R

#'--------------------------------------------------
#' Script for gmail analysis between two people
#' using gmailr
#'--------------------------------------------------

library(gmailr)
library(tidyverse)
library(rties)
library(lubridate)
library(gganimate)


#'-----------------------
## Connect to gmail  ####
#'-----------------------

## Here are some steps for doing this:
## 1. Go to https://console.developers.google.com/
## 2. Create a new project
## 3. Copy-and-paste the Client ID and Client Secret into the fields below


#'-----------------------
## authenticate  ####
#'-----------------------


client_id = "{YOUR CLIENT ID}"
client_secret = "{YOUR CLIENT SECRET}"
gm_auth_configure(key = client_id, secret = client_secret)

#'-----------------------
## get emails  ####
#'-----------------------

#this can be done better and in a more robust way. meh. it's a demo
from_lr <- gm_messages(search="from:[email protected] to:[email protected] before:2020/12/31 after:2019/12/31", num_results = 1e4)
from_me <- gm_messages(search="from:[email protected] to:[email protected] before:2020/12/31 after:2019/12/31", num_results = 1e4)


#'-----------------------------------------------------------
## a function to get email properties  ####
## that can be expanded in the future for other things
#'-----------------------------------------------------------
gm_get_properties <- function(input, a_string){
  
  #get raw messages
  msgs <- gm_id(input) %>%
    map(gm_message)
  
  #get dates from messages
  dates <- map_chr(msgs, gm_date)  %>% 
    dmy_hms()
  
  #other properties can go here....
  
  #assemble a return tibble
  return(tibble(from = a_string, dates = dates))
}

#'-----------------------------------------------------------
## Make dat data ####
#'-----------------------------------------------------------

email_dat <- bind_rows(gm_get_properties(from_lr, "lilarose"),
                       gm_get_properties(from_me, "me"))

#'-----------------------------------------------------------
## Plot the raw emails ####
## This can probably be done better
#'-----------------------------------------------------------

ggplot(email_dat,
       aes(x = dates, y = 1, color = from)) +
  geom_jitter(position = position_jitter(width = 0, height = 0.02)) +
  scale_y_continuous(limits = c(0.95,1.05))


#'-----------------------------------------------------------
## Make an hourly timseries of an email heartbeat ####
#'-----------------------------------------------------------


email_dat_summary <- email_dat %>%
  group_by(from,
           dates = floor_date(dates, "hour")) %>%
  summarize(emails = n())  %>%
  ungroup() 


#what? I liked this typo variable name
heartbeet <- email_dat_summary %>%
  group_by(from) %>%
  padr::pad(interval = "hour") %>%
  padr::fill_by_value(emails, value = 0) %>% 
  ungroup()


#an animation of the timeseries
ggplot(heartbeet %>% 
         filter(dates > floor_date(mdy("11/1/2020"))),
       aes(x = dates, y = emails, color = from)) +
  geom_line() +
  geom_point(size = 2) +
  labs(y = "# of emails", x = "", color = "from who") +
  scale_color_manual(values = c("red", "blue")) +
  theme_minimal() +
  theme(axis.text.x = element_text(angle=45, hjust = 1)) +
  transition_reveal(dates) 

anim_save("email_me_lr_series.gif")


heartbeet %>% 
  group_by(from) %>%
  arrange(dates) %>%
  mutate(cummulative_emails = cumsum(emails)) %>%
ggplot(aes(x = dates, y = cummulative_emails, color = from)) +
  geom_line() +
  geom_point(size = 2) +
  labs(y = "cummulative\n# of emails", x = "", color = "from who") +
  scale_color_manual(values = c("red", "blue")) +
  theme_minimal() +
  theme(axis.text.x = element_text(angle=45, hjust = 1)) +
  transition_reveal(dates) 

anim_save("email_me_lr.gif")