micahwoods/log2_twitter_followers_tweet_rate.R

## log2_twitter_followers_tweet_rate.R
# make a chart of tweets by followers and with log2 axes
# as shown in https://blog.codecentric.de/en/2017/07/combining-social-network-analysis-topic-modeling-characterize-codecentrics-twitter-friends-followers/
# and adapted from code at https://github.com/ShirinG/blog_posts_prep/blob/master/twitter/twitter_codecentric.Rmd

library(twitteR)
library(dplyr)
library(ggrepel)
library(cowplot)

api_key <- "your API_KEY here"

api_secret <- "your API_SECRET here"

access_token <- "your ACCESS_TOKEN here"

access_token_secret <- "your ACCESS_TOKEN_SECRET here"

setup_twitter_oauth(api_key, api_secret, access_token, access_token_secret)

user <- getUser("asianturfgrass")

friends <- user$getFriends() # who I follow
friends_df <- twListToDF(friends)
followers <- user$getFollowers() # my followers
followers_df <- twListToDF(followers)

atcFull <- followers_df

## # I want to get unique, also I would like to check on something, such as they follow
# both me and GCSAA; for that reason maybe it pulls out some of the non-turf accounts

# let me check gcsaa here
user2 <- getUser("gcsaa")
gcsaa <- user2$getFollowers() # gcsaa followers
gcsaa_followers_df <- twListToDF(gcsaa)

# let it be they must follow me and gcsaa or biggaltd
user3 <- getUser("BIGGALtd")
bigga <- user3$getFollowers() # bigga followers
bigga_followers_df <- twListToDF(bigga)

# and one more, AGCSA also
user4 <- getUser("AGCSA2")
agcsa <- user4$getFollowers()
agcsa_followers_df <- twListToDF(agcsa)

# to find the accounts that are probably really into the profession,
# subset to be only those that follow ATC, plus also follow GCSAA & BIGGA & AGCSA
atc_assoc_df <- subset(atcFull, screenName %in% gcsaa_followers_df$screenName &
                         screenName %in% bigga_followers_df$screenName &
                         screenName %in% agcsa_followers_df$screenName)

# find from all ATC followers who themselves have the most followers
top_fol <- atcFull %>%
  mutate(date = as.Date(created, format = "%Y-%m-%d"),
         today = as.Date("2017-07-30", format = "%Y-%m-%d"),
         days = as.numeric(today - date),
         statusesCount_pDay = statusesCount / days) %>%
  select(screenName, followersCount, statusesCount_pDay) %>%
  arrange(desc(followersCount)) %>%
  .[1:25, ]

# find from the subsetted followers who has the most followers
top_fol_assoc <- atc_assoc_df %>%
  mutate(date = as.Date(created, format = "%Y-%m-%d"),
         today = as.Date("2017-07-30", format = "%Y-%m-%d"),
         days = as.numeric(today - date),
         statusesCount_pDay = statusesCount / days) %>%
  select(screenName, followersCount, statusesCount_pDay) %>%
  arrange(desc(followersCount)) %>%
  .[1:25, ]

# find from all atc followers who tweets the most
top_tweet <- atcFull %>%
  mutate(date = as.Date(created, format = "%Y-%m-%d"),
         today = as.Date("2017-07-30", format = "%Y-%m-%d"),
         days = as.numeric(today - date),
         statusesCount_pDay = statusesCount / days) %>%
  select(screenName, followersCount, statusesCount_pDay) %>%
  arrange(desc(statusesCount_pDay)) %>%
  .[1:25, ]

# find from subsetted atc followers who tweets the most
top_tweet_assoc <- atc_assoc_df %>%
  mutate(date = as.Date(created, format = "%Y-%m-%d"),
         today = as.Date("2017-07-30", format = "%Y-%m-%d"),
         days = as.numeric(today - date),
         statusesCount_pDay = statusesCount / days) %>%
  select(screenName, followersCount, statusesCount_pDay) %>%
  arrange(desc(statusesCount_pDay)) %>%
  .[1:25, ]

# for all atc followers, combine the 25 with most followers
# and the 25 with the most tweets
top_fol_tweet <- rbind(top_fol, top_tweet)

# ensure that the screenName is not duplicated
top_fol_tweet <- top_fol_tweet[!duplicated(top_fol_tweet[ , 1]),]

# for atc followers who also follow AGCSA, BIGGA, and GCSAA, combine the 25 with most followers
# and the 25 with the most tweets
top_fol_tweet_assoc <- rbind(top_fol_assoc, top_tweet_assoc)

# ensure that the screenName is not duplicated
top_fol_tweet_assoc <- top_fol_tweet_assoc[!duplicated(top_fol_tweet_assoc[ , 1]),]

# function to make the decimal points less
axisDigits2 <- function(x) prettyNum(x, digits = 2)

full <- atcFull %>%
  mutate(date = as.Date(created, format = "%Y-%m-%d"),
         today = as.Date("2017-07-30", format = "%Y-%m-%d"),
         days = as.numeric(today - date),
         statusesCount_pDay = statusesCount / days) %>%
  ggplot(aes(x = followersCount, y = statusesCount_pDay)) +
    background_grid(major = "xy") +
    geom_smooth(method = "lm", se = FALSE, colour = "grey") +
    geom_point(color = "#3f7300", alpha = 0.4) +
    geom_text_repel(data = top_fol_tweet, aes(label = screenName), size = 3) +
    scale_x_continuous(trans='log2') +
    scale_y_continuous(trans='log2', labels = axisDigits2) +
    labs(x = expression(log[2]~"number of followers"),
         y = expression(log[2]~"average tweets per day"),
         title = "asianturfgrass's most influential followers",
         caption = "Second degree followers and tweet rate of @asianturfgrass Twitter followers (data from 30 July 2017)\nadapted from https://github.com/ShirinG/blog_posts_prep/blob/master/twitter/twitter_codecentric.Rmd")

full
save_plot("atc_followers.svg", full, base_width = 12, base_height = 6.75)

atc_assoc <- atc_assoc_df %>%
  mutate(date = as.Date(created, format = "%Y-%m-%d"),
         today = as.Date("2017-07-30", format = "%Y-%m-%d"),
         days = as.numeric(today - date),
         statusesCount_pDay = statusesCount / days) %>%
  ggplot(aes(x = followersCount, y = statusesCount_pDay)) +
  background_grid(major = "xy") +
  geom_smooth(method = "lm", se = FALSE, colour = "grey") +
  geom_point(color = "#3f7300", alpha = 0.4) +
  geom_text_repel(data = top_fol_tweet_assoc, aes(label = screenName), size = 3) +
  scale_x_continuous(trans='log2') +
   scale_y_continuous(trans='log2', labels = axisDigits2) +
  labs(x = expression(log[2]~"number of followers"),
       y = expression(log[2]~"average tweets per day"),
       title = "asianturfgrass's most influential followers",
       subtitle = "among those accounts that also follow GCSAA, BIGGALtd, & AGCSA2",
       caption = "Second degree followers and tweet rate of @asianturfgrass Twitter followers (data from 30 July 2017)\nadapted from https://github.com/ShirinG/blog_posts_prep/blob/master/twitter/twitter_codecentric.Rmd")

atc_assoc
save_plot("atc_followers_assoc.svg", atc_assoc, base_width = 12, base_height = 6.75)
	# make a chart of tweets by followers and with log2 axes
	# as shown in https://blog.codecentric.de/en/2017/07/combining-social-network-analysis-topic-modeling-characterize-codecentrics-twitter-friends-followers/
	# and adapted from code at https://github.com/ShirinG/blog_posts_prep/blob/master/twitter/twitter_codecentric.Rmd

	library(twitteR)
	library(dplyr)
	library(ggrepel)
	library(cowplot)

	api_key <- "your API_KEY here"

	api_secret <- "your API_SECRET here"

	access_token <- "your ACCESS_TOKEN here"

	access_token_secret <- "your ACCESS_TOKEN_SECRET here"

	setup_twitter_oauth(api_key, api_secret, access_token, access_token_secret)

	user <- getUser("asianturfgrass")

	friends <- user$getFriends() # who I follow
	friends_df <- twListToDF(friends)
	followers <- user$getFollowers() # my followers
	followers_df <- twListToDF(followers)

	atcFull <- followers_df

	## # I want to get unique, also I would like to check on something, such as they follow
	# both me and GCSAA; for that reason maybe it pulls out some of the non-turf accounts

	# let me check gcsaa here
	user2 <- getUser("gcsaa")
	gcsaa <- user2$getFollowers() # gcsaa followers
	gcsaa_followers_df <- twListToDF(gcsaa)

	# let it be they must follow me and gcsaa or biggaltd
	user3 <- getUser("BIGGALtd")
	bigga <- user3$getFollowers() # bigga followers
	bigga_followers_df <- twListToDF(bigga)

	# and one more, AGCSA also
	user4 <- getUser("AGCSA2")
	agcsa <- user4$getFollowers()
	agcsa_followers_df <- twListToDF(agcsa)

	# to find the accounts that are probably really into the profession,
	# subset to be only those that follow ATC, plus also follow GCSAA & BIGGA & AGCSA
	atc_assoc_df <- subset(atcFull, screenName %in% gcsaa_followers_df$screenName &
	screenName %in% bigga_followers_df$screenName &
	screenName %in% agcsa_followers_df$screenName)

	# find from all ATC followers who themselves have the most followers
	top_fol <- atcFull %>%
	mutate(date = as.Date(created, format = "%Y-%m-%d"),
	today = as.Date("2017-07-30", format = "%Y-%m-%d"),
	days = as.numeric(today - date),
	statusesCount_pDay = statusesCount / days) %>%
	select(screenName, followersCount, statusesCount_pDay) %>%
	arrange(desc(followersCount)) %>%
	.[1:25, ]

	# find from the subsetted followers who has the most followers
	top_fol_assoc <- atc_assoc_df %>%
	mutate(date = as.Date(created, format = "%Y-%m-%d"),
	today = as.Date("2017-07-30", format = "%Y-%m-%d"),
	days = as.numeric(today - date),
	statusesCount_pDay = statusesCount / days) %>%
	select(screenName, followersCount, statusesCount_pDay) %>%
	arrange(desc(followersCount)) %>%
	.[1:25, ]

	# find from all atc followers who tweets the most
	top_tweet <- atcFull %>%
	mutate(date = as.Date(created, format = "%Y-%m-%d"),
	today = as.Date("2017-07-30", format = "%Y-%m-%d"),
	days = as.numeric(today - date),
	statusesCount_pDay = statusesCount / days) %>%
	select(screenName, followersCount, statusesCount_pDay) %>%
	arrange(desc(statusesCount_pDay)) %>%
	.[1:25, ]

	# find from subsetted atc followers who tweets the most
	top_tweet_assoc <- atc_assoc_df %>%
	mutate(date = as.Date(created, format = "%Y-%m-%d"),
	today = as.Date("2017-07-30", format = "%Y-%m-%d"),
	days = as.numeric(today - date),
	statusesCount_pDay = statusesCount / days) %>%
	select(screenName, followersCount, statusesCount_pDay) %>%
	arrange(desc(statusesCount_pDay)) %>%
	.[1:25, ]

	# for all atc followers, combine the 25 with most followers
	# and the 25 with the most tweets
	top_fol_tweet <- rbind(top_fol, top_tweet)

	# ensure that the screenName is not duplicated
	top_fol_tweet <- top_fol_tweet[!duplicated(top_fol_tweet[ , 1]),]

	# for atc followers who also follow AGCSA, BIGGA, and GCSAA, combine the 25 with most followers
	# and the 25 with the most tweets
	top_fol_tweet_assoc <- rbind(top_fol_assoc, top_tweet_assoc)

	# ensure that the screenName is not duplicated
	top_fol_tweet_assoc <- top_fol_tweet_assoc[!duplicated(top_fol_tweet_assoc[ , 1]),]

	# function to make the decimal points less
	axisDigits2 <- function(x) prettyNum(x, digits = 2)

	full <- atcFull %>%
	mutate(date = as.Date(created, format = "%Y-%m-%d"),
	today = as.Date("2017-07-30", format = "%Y-%m-%d"),
	days = as.numeric(today - date),
	statusesCount_pDay = statusesCount / days) %>%
	ggplot(aes(x = followersCount, y = statusesCount_pDay)) +
	background_grid(major = "xy") +
	geom_smooth(method = "lm", se = FALSE, colour = "grey") +
	geom_point(color = "#3f7300", alpha = 0.4) +
	geom_text_repel(data = top_fol_tweet, aes(label = screenName), size = 3) +
	scale_x_continuous(trans='log2') +
	scale_y_continuous(trans='log2', labels = axisDigits2) +
	labs(x = expression(log[2]~"number of followers"),
	y = expression(log[2]~"average tweets per day"),
	title = "asianturfgrass's most influential followers",
	caption = "Second degree followers and tweet rate of @asianturfgrass Twitter followers (data from 30 July 2017)\nadapted from https://github.com/ShirinG/blog_posts_prep/blob/master/twitter/twitter_codecentric.Rmd")

	full
	save_plot("atc_followers.svg", full, base_width = 12, base_height = 6.75)

	atc_assoc <- atc_assoc_df %>%
	mutate(date = as.Date(created, format = "%Y-%m-%d"),
	today = as.Date("2017-07-30", format = "%Y-%m-%d"),
	days = as.numeric(today - date),
	statusesCount_pDay = statusesCount / days) %>%
	ggplot(aes(x = followersCount, y = statusesCount_pDay)) +
	background_grid(major = "xy") +
	geom_smooth(method = "lm", se = FALSE, colour = "grey") +
	geom_point(color = "#3f7300", alpha = 0.4) +
	geom_text_repel(data = top_fol_tweet_assoc, aes(label = screenName), size = 3) +
	scale_x_continuous(trans='log2') +
	scale_y_continuous(trans='log2', labels = axisDigits2) +
	labs(x = expression(log[2]~"number of followers"),
	y = expression(log[2]~"average tweets per day"),
	title = "asianturfgrass's most influential followers",
	subtitle = "among those accounts that also follow GCSAA, BIGGALtd, & AGCSA2",
	caption = "Second degree followers and tweet rate of @asianturfgrass Twitter followers (data from 30 July 2017)\nadapted from https://github.com/ShirinG/blog_posts_prep/blob/master/twitter/twitter_codecentric.Rmd")

	atc_assoc
	save_plot("atc_followers_assoc.svg", atc_assoc, base_width = 12, base_height = 6.75)