Get twitter conference coverage based on a hashtag over time with R twitteR package.

Visualization made with ggplot2, scales and wordcloud packages

./conferenceCoverage.R -a HASTAG -d DATE -l LENGTH [-f MINFILTER -u MAXUSERS -t MAXTWEETS]

• HASTAG is the hastag for the conference (without #)
• DATE is the first day of the conference in YYYY-MM-DD format
• LENGTH is the length in days of the conference
• MINFILTER is the minimum number of tweets/user in the top user graph [default is 3]
• MAXUSERS is the maximum number of users in the top user graph [default is 40]
• MAXTWEETS is the maximum number of tweet/day required by the API [default is 1500]

./conferenceCoverage.R -a JOBIM2016 -d 2016-06-26 -l 5
./conferenceCoverage.R -a JOBIM2016 -d 2016-06-26 -l 5 -f 5 -u 25 -t 1000

In Shell:

sudo apt-get install r-base r-base-dev libxml2-dev libcurl4-openssl-dev curl libcairo-dev

If problems with svglite, check github.com/hadley

In R console:

install.packages("devtools")
install.packages("RCurl")
install.packages("twitteR")
install.packages("ggplot2")
install.packages("tm")
install.packages("scales")
install.packages("wordcloud")
install.packages("optparse")

• Log to your account on twitter.com (or create a new one)
• Go to dev.twitter.com
• Create an app to get a consumerKey and a consumerSecret

In a R console :

library(twitteR)  
consumer_key <- 'YOURCONSUMERKEY'
consumer_secret <- 'YOURCONSUMERSECRET'
access_token <- 'YOURACCESSTOKEN'
access_secret <- 'YOURACCESSSECRET'
save(list = c("consumer_key", "consumer_secret", "access_token", "access_secret"), file="cred")

The cred file just created contain all your credentials for Twitter, so don't share it on Github ;-). Now you just had to load it and you can authentify yourself with it.

• Add more stuff:
  • Choice of color theme
  • Shiny

Script is mostly copied and slightly adapted on Neil and Stephen twitter analysis:

• github.com/neilfws
• github.com/stephenturner

Informations for installing twitteR and RCurl on Ubuntu:

• freakonometrics