Twitter is one way that I keep up to date on Great Lakes news. Might I be missing out on some good Great Lakes tweeters? I used the R package rtweet to investigate (see this vignette).
library(rtweet)
library(tidyverse)
library(RColorBrewer)
library(lubridate)
# look at the last ntwee
ntweets <- 18000
gltweets <- search_tweets('
"#GreatLakes" OR "Great Lakes" OR
"#LakeSuperior" OR "Lake Superior" OR
"#LakeMichigan" OR "Lake Michigan" OR
"#LakeHuron" OR "Lake Huron" OR
"#LakeErie" OR "Lake Erie" OR
"#LakeOntario" OR "Lake Ontario"',
n=ntweets, include_rts=FALSE, verbose=FALSE)
goback <- signif(diff(range(gltweets$created_at)), digits=2)
First, I looked at the last 18,000 tweets that mentioned the Great Lakes. These tweets go back 8.1 days.
Then I counted up the number of Great Lakes tweets for each tweeter.
gltweeters <- gltweets %>%
users_data() %>%
count(screen_name, name, description, followers_count) %>%
arrange(-n, -followers_count) %>%
select(No.GL.tweets=n, Handle=screen_name, Name=name,
Description=description, No.followers=followers_count)
library(DT)
datatable(gltweeters, rownames=FALSE)
Not as enlightening as I had hoped. I wanted to standardize the number of Great Lakes tweets by the total number of tweets from each person over the same time period, but I couldn’t figure out how to do that.