`What are you Tweeting about?':
A survey of Trending Topics within Twitter
Marc Cheong
Clayton School of Information Technology,
Monash University
Victoria, 3800
Australia
marc.cheong@infotech.monash.edu.au
Abstract. Twitter allows users to observe the top ten popular terms or
topics of discussion at any given moment through its `Trending Topics'
feature. In this paper, we monitor the Trending Topics chart to survey
the topics frequently discussed within the Twitter community (i.e. the
`Twitterverse'). This allows us to gain better insight into the collective
viewpoint and zeitgeist exhibited by the Twitter ecosystem as a whole;
and also to learn more about the subjects of interest of the typical Twit-
ter user by looking at the `big picture' context of the common types of
Twitter messages that are promoted to Trending Topic status.
1 Introduction
Twitter [13] has evolved from being a simple microblogging service with the sole
objective of answering the trivial question \what are you doing?" to an Web 2.0
phenomenon that has a large user base with exponential growth rates [18]. It is
now being used globally by people from all walks of life, including celebrities,
politicians, and organizations [2].
One of the interesting features of Twitter is the existence of Trending Topics,
which is a list of top ten most tweeted topics ranked by Twitter's proprietary
algorithm (`Tweeting' is a term for writing Twitter messages). This list is readily
available at http://twitter.com/ as can be seen in Figure 1.
Fig. 1. A screenshot of Twitter Trending Topics accessible via http://twitter.com/.

Trending Topics are a useful tool to discover the topics of interest to the
`Twitterverse' { a term used by Twitter users to describe the Twitter community
as a whole, similar to the usage of the term blogosphere for blogs { as a whole. In
other words, it is a simple way of measuring the collective and emergent behavior
exhibited by Twitter users worldwide at any given moment.
The motivation behind this paper is to survey the list of keywords present
on Trending Topics, study the signicance these topics have on the Twitter user
base, and observe how these topics evolve over time. This ideally gives us a `big
picture overview' to better understand the nature of information that can be
mined from the Twitterverse, and also the nature of the latent demographics
behind the Twitterverse itself.
2 Related Work
As academic research on Twitter is generally a new eld, there exists a limited
amount of research into the Twitter ecosystem.
Specically, discussion on Twitter Trends is limited to Cheong & Lee [2]
who analyzed the demographics behind a sampling of both trending and non-
trending topics by discovering the inherent user base and manually identifying
properties of the users as discovered via the Twitter API [13]. They have found
that by retrieving the messages discussing about a particular trend, the user
demographics and Twitter usage patterns exhibited will mirror the subject's
real-world properties. As an example, they determined that the Trending Topic
Grey's Anatomy (a US television drama) has a Twitter user demographic that
accurately resembles the real-world viewership of the TV show [2].
The key dierence between this paper and [2] is that Cheong & Lee have
performed a `top-down' analysis detailing the specics of the user base (demo-
graphics, usage habits) of a particular trend. In this paper, however, we per-
form an overview of the Trending Topics themselves { a high-level, `big picture'
overview { without assuming anything about the underlying user base.
On a broader scope, literature on various aspects of Twitter in academic re-
search has been growing within the past two years. Krishnamurthy et al. [12]
has performed research on the growth rates, geographic spread, and other sta-
tistical properties on the Twitter user base. Java et al. [11] and Huberman et
al. [9] performed similar research, but with more emphasis on the social network
perspective of Twitter. To a certain extent, Java et al. [11] has worked on an-
alyzing popular keywords on Twitter by focusing on phrase rankings restricted
to a particular Twitter user network (composed of a subset of users interlinked
with each other via a follow connection on Twitter).
A focused usage of Twitter on emergency response and mass convergence {
Twitter usage habits during hurricanes, and US presidential campaigns, respec-
tively { has been studied by Hughes & Palen [10]. A prototype framework of
using Twitter to enable authorities to better respond to acts of terror has been
proposed by Cheong & Lee [3]. Such uses of Twitter are given consideration in
our review of prior work, as crisis topics and topics of mass convergence (as clas-
2