Title: Extracting Information from Social Media Using R
Abstract: In this short talk we will present recent work with Luciana Dalla Valle and Andrea Baldino
on using R to extract information from Facebook and Twitter. We will describe some
applications of the Rfacebook package written by Pablo Barbera and Michael Piccirilli
and of the twitteR package written by Jeff Gentry. For Facebook we will illustrate how
the average number of likes, comments and shares across posts to a page change with
time. We will discuss some of the character manipulation code that we have developed to
clean and process posts. We will then see the graphical presentation of word counts using
bar plots and word clouds. We will also outline a simple form of Sentiment Analysis and
the presentation of its results. We finish our treatment of the analysis of data available
from Facebook by displaying the friends of a user as a network.
At the time of writing, Twitter allows us to download a somewhat richer data set than
Facebook. As with Facebook, we can obtain word counts, produce word clouds and
perform Sentiment Analyses for sets of tweets. We can also assess how the number of tweets
varies across day and time. We can extract information for individual users including
their followers and those whom they are following. We finish by illustrating the results of
a cluster analysis on this set of followers and followed.
Depending on the bravery of the speaker, some of these techniques may be demonstrated
live using R.