Identity Use and Misuse of Public Persona on Twitter
Dicle Berfin Köse, Jari Veijalainen and Alexander Semenov
Department of Computer Science and Information Systems, University of Jyvaskyla,
FI-40014, Jyvaskyla, Finland
Keywords: Twitter, Impersonation in Social Media, Faked Accounts, Online Identity, G20 Leaders, Putin, Erdogan,
Obama.
Abstract: Social media sites have appeared during the last 10 years and their use has exploded all over the world.
Twitter is a microblogging service that has currently 320 million user profiles and over 100 million daily
active users. Many celebrities and leading politicians have a verified profile on Twitter, including Justin
Bieber, president Obama, and the Pope. In this paper we investigate the '‘hundreds of Putins and Obamas
phenomenon’ on Twitter. We collected two data sets in 2015 containing 582 and 6477 profiles that are
related to the G20 leaders’ profiles on Twitter. The number of namesakes varied from 5 to 1000 per leader.
We analysed in detail various aspects of the Putin and Erdogan related profiles. For the first ones we looked
into the language of the profiles, their follower sets, the address in the profile and where the tweets were
really sent from. For both profile sets we investigated why the accounts were created. For this, we deduced
12 categories based on the information in the profile and the contents of the sent tweets. The research is
exploratory in nature, but we tentatively looked into online identity, communication and political theories
that might explain emergence of these kinds of Twitter profiles.
1 INTRODUCTION
Many social media sites have been created during
the last 10 years and Facebook has now over one
billion users and many other sites have hundreds of
millions of users. The Chinese microblog service
Sina Weibo (China, 2015), for instance, has over
500 million users mainly from the mainland China.
The first microblog service in the world, Twitter
(Twitter, Inc., 2005), has 320 million monthly active
users at the time of writing. 79% of the users of
Twitter have indicated that their address is outside of
USA and 80% of them use mobile terminals to
access Twitter. The site currently supports 35+
languages, including practically all European
languages, Japanese, Korean, Hindi, Arabic, Persian
and Chinese. 500 million tweets are submitted daily
to the site.
Twitter offers 140 character long messages,
tweets, that users can create using a browser or a
smart phone application. After uploading it to the
site, the tweet is distributed to the followers of the
user. All the tweets are stored by the site. If the
tweet is public, it can be found by the search
engines. In the site-internal search engine it is
possible to use any string as a keyword, including
hashtags (#...). The followers will get the tweets
once they have logged in to the site. A user can
retweet a tweet to his or her followers. In this way a
tweet can reach a much larger set of users than the
originator of the tweet has. One can also send a
private tweet to user by using his or her screen name
as the sole address at the beginning. Mentioning
user’s screen name inside a tweet notifies that user
as well, even if he or she is not a follower.
Twitter also has APIs through which the public
tweets as well as the complete user profile data can
be retrieved. The followers of a user are also
retrievable. We will use these features in this study.
The paper is organized as follows. In Section II
we introduce theories that are helpful in interpreting
the ‘hundreds of Putins and Obamas phenomenon’
in Twitter. In section III we will describe the data set
we are using in our analysis. In section IV we will
present the results concerning the deduced profile
categories. Section V contains a short related work
part and VI concludes.
164
Köse, D., Veijalainen, J. and Semenov, A.
Identity Use and Misuse of Public Persona on Twitter.
In Proceedings of the 12th International Conference on Web Information Systems and Technologies (WEBIST 2016) - Volume 1, pages 164-175
ISBN: 978-989-758-186-1
Copyright
c
 2016 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved

2 THEORIES ABOUT SOCIAL
MEDIA PRESENCE AND
IMPERSONATION
2.1 Some Preliminary Observations on
the Data
We first share some observations that give us clues
about possible reasons for the existence thousands of
accounts that are related with public figures. The
Russian president has two verified profiles,
‘President of Russia @KremlinRussia_E’ (with
user_id 205622130 and 296598 followers in March
2015), established in October 2010 in English, and
the Russian profile ‘Президент России
@KremlinRussia’ (with user_id 158650448 and
2139735 followers in March 2015), established in
June 2010. It is obvious that president Putin does not
have a need to establish almost 600 (actually almost
800 in Nov. 2015) profiles in Twitter for himself
under his name or under alias, although he might
have an account carrying his own name. None of
those almost 600 matching accounts we found is
verified, except the above two. Thus, we do not
know whether president Putin runs an account of his
own, under his name or under some other identity in
Twitter or not. Three accounts claim to be run by
Putin, @PutinRF_Eng had over 225000 followers. It
joined Twitter in Nov. 2012 and has been active
until the last days. Another account with the screen
name @PutinRF has over 1 million followers and it
also says that it is an official account of Putin. This
account mostly tweets in Russian and has joined
Twitter in Dec. 2011. @PutinRF_Ita also exists, but
it only has tweeted a short time. Further, there is an
account in Arabic with screen name
@Vladimirarabia, but it has been active only a short
time in Sept. 2011. In the profile it claims to be
controlled by president Putin, as well. A fifth
account is @PutinRF2012 that joined in Dec. 2011
but that account has been passive since February
2012. It is claimed in the profiles of @PutinRF that
if Putin himself tweets, the tweet is signed by
VP/BП. In any case, it is highly probable that the
majority of the hundreds of accounts carrying his
name in the name field or in the screen name of the
profile are established and controlled by other
people, partially also outside Russia. Some of the
accounts announce in the profile that they are
parody, commentary or fan accounts according to
Twitter rules (Twitter Help Center, 2014), but most
do not belong to these categories. Thus, it is possible
that they are impersonating president Putin in the
sense Twitter defines it – or the people controlling
them want to hide their true identity by using Putin’s
identity.
Looking at Obama-related accounts we can
observe similar phenomena. There are almost 1000
accounts in our data set that are related with Obama.
Some accounts indicate clearly that they are parody
accounts or fan/supporter accounts. There are also
accounts that claim to be controlled by president
Obama, but most probably are not. President Obama
only has two verified accounts, @BarackObama and
@POTUS. The former was created as early as in
March 2007, the latter in June 2013. President
Obama is controlling both, but mostly his aides
tweet through the former (unless tweet is signed by –
BO). As an example of a confusing profile one can
pick one with the screen name @President and name
‘US President News’. It is not verified and it also
states in the bio to be unofficial, although it carries
the official seal of the US president in its profile
picture. It was created in March 2011, has circa 50
thousand followers and has tweeted over 60000
times, i.e. tens of tweets per day in average. From
the contents of the tweets and URLs it often includes
into the tweets one can conclude that it is critical
about Obama’s politics, while keeping meticulously
track of his appearances and statements. To better
understand what might be the reasons behind
creating profiles that utilize the famous leader’s
identity to a smaller or greater extent, we first look
at theories that might be of relevance in explaining
the phenomena.
2.2 Anonymity
Anonymity is the situation where the message
source is unknown or it is hidden to a large extent
(Scott, 2004). That means that the person, i.e. the
unique biological human being sending the message,
is not identified by others (Lapidot-Lefler and
Barak, 2012).
Staying anonymous is dependent on and in
relation to a certain context and medium (Suler,
2002). On the Internet, the amount of personal
information given may be chosen by the individual;
therefore, online identity may be between true
anonymity and fully identified (Ardia, 2012). In the
latter case the digital identity (such as name, picture,
social security number, used address, etc.) used by
the communicating human being can be traced back
to him or her with certainty. It is also dependent on
the online service used (Zhao et al., 2008), and to
which extent it allows its users control their social
presence through employment of various identity
Identity Use and Misuse of Public Persona on Twitter
165

cues (Lapidot-Lefler and Barak, 2012). For instance,
Twitter has the information using which IP address
the profile @PutinRF_Eng was created and from
which IP addresses it is being used. But it does not
necessarily know the real name of the person(s)
issuing the tweets. The users following the profile do
not even know the IP-addresses. If the account is
verified, though, Twitter Inc. guarantees that the real
person or organisation identified on the profile has
indeed created (or later rendered control over) the
profile and the issued tweets originate from this
identified source. This is a strong identity cue.
Online anonymity influences how the Internet is
used. The major effects may be listed as online
disinhibition effect, enabling and encouraging free
expression, changes in the quality and quantity of
comments, and exploitation of the Internet for
malicious activities.
Suler (2004) defines online disinhibition affect as
the less restrained behaviour on cyberspace
compared to face-to-face communication. It has two
types: toxic disinhibition and benign disinhibition.
Toxic disinhibition implies usage of offensive
language, cruel comments, or surfing criminal
websites (Suler, 2004), and it usually damages other
people’s images (Lapidot-Lefler and Barak, 2012).
Benign disinhibition stands for the salutary effects
that may cover sharing intimate information, or acts
of kindness or generosity (Suler, 2004); and may
involve self-therapeutical effects through increased
amount of confessional self-disclosures (Belk,
2013).
Anonymity provides a safe haven for those who
are afraid to disclose their identities when expressing
their views (Santana, 2014). Therefore, it increases
speech variety (Akdeniz, 2002), and encourages
exchange of different types of information and
opinion (Kaye, 2015). Furthermore, it influences
participating in processes of social and political
change (Hollenbeck and Zinkhan, 2006), hence it is
essential in repressive regimes. To that effect, online
social networks provide disguise for pro-democracy
activists and journalists (Bodle, 2013).
However, recent studies (Santana, 2014;
Fredheim et al., 2015) show that anonymity
decreases the level of civility of online discussions,
and when their identity is known people tend to
comment less, pay attention not to make typos,
avoid obscene language and shift their remarks from
personalities to issues.
Unfortunately, anonymity enables exploitation of
the Internet for malicious purposes, as well.
Spamming, deception, hate mailing, impersonation
and mispresentation, online financial fraud
(Christopherson, 2007; Kling et al., 1999), cyber-
smearing, flaming, online terrorist activities, various
forms of cybercrime such as high tech paedophilia,
difficulty of credibility evaluation on important
issues, and inability to get credit for input/ideas
especially in decision-making systems (Scott, 2004)
may be listed as the problems anonymity creates on
cyberspace.
2.3 Identity and Impression
Management
As a socially constructed concept, identity differs
from the sense of self because, it is the way the self
is known to others, and it requires existence of other
people (Altheide, 2000). Since how the identity is
perceived by others affects the way the person is
treated, individuals try to control their impressions
on people. Goffman (1959) calls this management of
identity impression management. Leary and
Kowalski (1990) define impression management as
the behavioural attempts to influence the perception
of others about ourselves. Dividing their identity
into two, private and public; individuals adjust their
public identity according to the situation by playing
conditional characters so that they appear attractive
to people surrounding them, and they use their
private identity as a preparation phase for their
public performance. (Goffman, 1959.) According to
Miller (1995) online communication provides a new
platform for self-presentation through assertions and
displays about the person. Especially online social
networks enable their members to, in Sundén’s
words (2003), “type oneself into being”. Online,
people may show various features of their identity
without the obligation to fully present themselves
(Suler, 2002).
Yet, the ubiquitous nature of the Internet, and the
relative lack of control on the audience necessitate
that online identities are kept under control. And it is
even more important for celebrities to do so because
of the commercial value of their identity and the
constant public scrutiny on them. The online
identities of famous people are a continuum of their
branded-selves and should continue to attract
attention and to acquire cultural and monetary value
(Marshall, 2010; Hearn, 2008). Politicians use their
online identities to communicate with voters and to
make political statements without any intermediary
media (Skogerbø and Krumsvik, 2015). Their
characters displayed on media convey their values,
and eventually this influences how their policies are
perceived by the citizens, and whether or not the
public vote for them (Castells, 2007; Marshall,
WEBIST 2016 - 12th International Conference on Web Information Systems and Technologies
166

2010). A unique example of how politicians use
their online presences is Barack Obama’s 2008
presidential campaign, during which he used the
Internet for organizing voters, fund raising, and
advertisement (Kiss, 2008; Miller, 2008). The
profile @BarackObama established in March 2007
was a part of the campaign Artists use their online
presences as a continuum of their cultural merits
along with their main art form (Marshall, 2010).
3 CASE STUDIES
3.1 Data Set Description
Our data set consists of semi-manually collected
accounts from Twitter that are related with G20
leaders. The presidents or kings were selected from
the nation states, unless they are ceremonial figures,
like the German ‘Bundespresident’ and the Queen
Elizabeth from UK. From EU Donald Tusk and from
the European Central Bank the Director General
Mario Draghi was included. From Russia data also
for Prime Minister Dmitry Medvedev was collected
The collection was performed in two stages. In
March-April 2015 the search engine of Twitter was
used in order to find all users with the keywords like
*@Putin*, putin, or Путин. The searches returned
over 600 users. Some of them are just quoting the
word ‘Putin’ or Путин in their tweets, but almost
600 have the character string ‘Putin’, or a specific
modification of it (e.g. ‘putin’ or ’PUTIN’, or
‘Путин’) included into the screen name or into the
name field of the account. We have selected to our
data set 579 accessible accounts where the character
string ‘putin’, or a modification where one or more
of the letters have been replaced by a corresponding
capital letter, appears as part of the screen name. In
addition, we selected all accounts where Putin or
some modification of it as above or Путин (or a
similar modification with capital letters as above)
appears in the name field. Among those 579 most
carried ‘putin’ or ‘vladimir putin’ with some
additions or omissions in their publicly accessible
screen name, such as @putinkgb or
@Putin_Vladimir. With these selection criteria we
have also accounts in our data set that are not really
related to president Putin, but a vast majority of
them are. We also found circa 10 accounts where a
slightly different screen name leads to the same
user_id and account inside Twitter. Such an account
is included only once into our data set. Three
accounts had been deleted or blocked and could not
be accessed at all. In addition to the 579 accounts
above, we included the accounts of president Obama
with screen name @BarackObama (user_id 813286
and circa 57.4 million followers), as well as two
verified accounts of the Russian prime minister,
Dmitry Medvedev, @MedvedevRussia (with user_id
153812887 and 3638691 followers) in Russian and
@MedvedevRussiaE (with user_id 153810519 and
914990 followers) in English. Thus, the entire data
set contained 582 profiles. We also collected all
followers of all those 582 profiles.
The second collection was performed in Nov.-
Dec. 2015 directed towards the leaders of G20,
including Putin. We have used Twitter API function
users/search, with full name of country leader as a
parameter. In the latter collection we found 786
Putin related profiles. Limitation of this approach is
that users/search returns only 1000 results.
We report here the results mainly from the earlier
collection for Putin, Medvedev and Obama. We did
not collect followers for all the 6477 profiles in Nov.
Dec. 2015.
Table 1: Number of Twitter Accounts for G20 Leaders.
Public Persona # Related Profiles
Cristina Fernandez de Kirchner 5
Malcolm Turnbull 54
Dilma Rousseff 332
Justin Trudeau 168
Xi Jinping 64
Francois Hollande 215
Angela Merkel 308
Narendra Modi 1000
Matteo Renzi 111
Shinzo Abe 46
Park Geun-hye 51
Enrique Pena Nieto 320
Vladimir Putin 786
king Salman 559
Jacob Zuma 153
David Cameron 999
Barack Obama 997
Donald Tusk 72
Mario Draghi 36
Recep Tayyip Erdogan 201
TOTAL 6477
Identity Use and Misuse of Public Persona on Twitter
167

3.2 Analysis of Putin Related Profiles
(Spring 2015 Collection)
As was discussed above, Russian president has two
officially verified profiles and some further profiles
might be controlled by him or by his aides. One
indication of the real controlling entity might be
other leaders of Russia and leaders of other big
countries. Interestingly, president Obama’s official
account followed @Putin, @president_putin, and
@VladimirPutin, and both official accounts of the
Prime Minister Medvedev above, but none of the
above verified accounts of the Russian president.
@KremlinRussia_E followed @BarackObama,
though. As can be expected, the prime minister’s
account followed @KremlinRussia and
@KremlinRussia_E, each other, and the latter also
follows @BarackObama. None of the prime
minister’s accounts follows, though, any other of the
accounts in our first data set, especially none of
those three “Putins” that president Obama’s account
follows.
It is clear that most of the profiles referring to
president Putin either by name in the name field or
in the screen name is not controlled by him or his
aides. This is because the content is often critical of
him or ridiculing him and some profiles tweet in
languages that are of less importance for the Russian
president. Some non-official profiles also only have
a few tweets and the last tweet was issued years
back. Most profiles tweet in English. Some profiles
announce their location, to be in Kremlin, in Russia,
or in other countries, like in USA, or UK. Most do
not indicate location at all. Thus, our assumption is
that with a high probability president Putin has a few
accounts in Twitter that he or his aides are
controlling, but vast majority are not controlled by
him or his aides.
3.3 User Profile Data (Spring 2015
Collection)
We collected the complete profile for each user in
the above data set of 579 profiles and those of
Medvedev and Obama. This is possible, even if the
tweets are protected. There were 22 profiles where
the tweets were protected. We parsed the user profile
data in order to get values for certain attributes. The
attributes in the user profile records are mostly
named in a similar manner as below. We have taken
the earliest point of time found in the user profile
record to mark the creation time and the latest point
of time anywhere in the record as the last activity
time. The times are given in the records with UTC
0000 offset, and with the resolution of one second,
but we only use the resolution of a day in our study.
From the user profile records we extract the
following information:
 user_id (id): what is the Twitter-internal unique
identifier for the screen_name
 account_created_at: the smallest timestamp
inside the record in any
created_at=datetime.datetime() item
 tweets_protected: (protected; True/False)
 language (lang): language of the account (e.g.
‘ru’, ‘en’)
 location (location); claimed location of the user
 followee_count (friends_count); number of
other profiles followed by the profile ,
 followers_count (followers_count); how many
followers the profile has?
 number_of_tweets (statuses_count); how many
tweets the profile has sent
 last_activity_at: the highest timestamp in the
created_at=datetime.datetime() item
The language attribute ‘lang’ can have many
values in a user record. We observed that there
might be at most two different in our data set. The
most often occurring is recorded as the language of
the profile. We found 274 profiles where the
language was Russian (ru), and 251 where the
language was English (en). Spanish (es) was
recorded for 16 profiles. Seven profiles were
categorized to use Japanese (ja). Major European
languages occurred as main language on few profiles
for each language. In five cases we could not
determine the language from the user profile record
automatically.
Of the accounts in our data set, circa 300 have
tweeted during March 2015 and later. About 240
have not tweeted during 2015 and 87 have not been
active since 2012.
3.4 Follower Data (Spring 2015
Collection)
For all 582 profiles we collected all the followers we
could. There were roughly 66.9 million of them. In
Table 1 all profiles with more than 10000 followers
are shown in the order of ascending creation dates. If
we ignore those with over 100000 followers the rest
of the profiles have at most 60000 followers and 97
% of the profiles have less than 10000 followers.
Among them the average number of followers is
654. However, if we drop only the 5 verified profiles
that all have more than 100000 followers, the
average number of followers jumps to 4230. The
overall number of followers of all non-verified
WEBIST 2016 - 12th International Conference on Web Information Systems and Technologies
168

profiles is 2444387, the number of distinct followers
is 2159743, and the overall number of distinct
followers except those of Obama is 5897322. The
overall number of distinct followers in the whole set
is 62496330. Thus, we can infer that almost 6
million users follow either the four verified profiles
controlled from Kremlin or the various unofficial
‘Putin’ profiles and circa 2.16 million follow at least
one profile in the latter category.
We have calculated the pairwise intersecting
follower sets for all users in our data set. The goal is
to investigate the distribution of the followers among
the profiles and also find out the reciprocal follower
relationships. The entire follower intersection table
contains 169026 rows. 82 % of the intersections are
empty. It is to be expected that those intersections
will be largest where both follower sets are among
the largest ones. Interestingly, only
@MedvedevRussia and @KremlinRussia have 1.43
million common followers, all other pairs have less
than one million. Only 15 profile pairs have more
than 100000 common followers, the pairs consisting
of all the verified five profiles, PutinRF, and
PutinRF_Eng (cf. Table 2).
As is usual in social media graphs, the
distributions are strongly skewed. This also holds for
the intersection sizes. There are circa 60
intersections, where the intersection size is over 50
% of the smaller follower set.
Table 2: Profiles with more than 10000 followers.
4 PROFILE CATEGORISATION
4.1 Profile and Tweets Types in the
Putin Related Set (Spring 2015
Collection)
The analysis of Vladimir Putin related profiles on
Twitter showed that there are clearly different kinds
of profiles. They were created for various reasons
including parody, impersonation, providing news,
using Putin’s identity for advertisement, political
campaigns, commentary; and some of them were
stated as bots.
Therefore, it was deemed appropriate to classify
the profile set according to its nature. The
categorization was established according to the
information on profile bios, and the content of the
tweets. It was assumed that the first tweet would
state the purpose of establishing the profile. Fig.1
below shows the deduced categories and the number
of profiles in each category.
Two verified profiles were found related to
Vladimir Putin: KremlinRussia and
KremlinRussia_E. Both profiles work as newsfeed
from Kremlin informing their followers about the
deeds of Vladimir Putin; most tweets link to
http://en.kremlin.ru website.
Personal profiles are the profiles that have
‘Putin’ in their name or screen name, but are not
pertained to Vladimir Putin. They may be
namesakes; among them onetimes that are set up to
post a tweet, or communicate without revealing the
real user identity, and usually used only once or
occasionally when required; or profiles which use
Putin’s name to hide the user identity or joke, but do
not post anything related to Vladimir Putin himself.
In total 295 personal profiles were found; 97 of them
were namesakes, one was a onetime profile, and 197
were using, Vladimir Putin’ as part of their digital
identity thus prohibiting mapping from digital
identity in Twitter to their real identity.
Adverts are the profiles that use Vladimir Putin’s
identity to attract followers, and post messages
related to their own promotion; or profiles that are
used to increase number of retweets or mentions of a
particular user. 19 profiles were found to be of this
nature.
Newsfeeds, as the name implies, are the ones
formed for objective of broadcasting reports or other
information. In total 23 profiles were found, 16 of
which were linked to a website or programme. One
of them is @putinizer with the highest degree in Fig.
3. It is related to http://putin.trendolizer.com/.
Another with protected tweets is @putinism_net that
Screen name Followers Created Lang
BarackObama 57467473 20070305 en
Putin 42571 20080219 en
Puitn_Vivat 20366 20090914 ru
putin2012 25153 20091012 ru
iPutin 13422 20100202 en
MedvedevRussia 3638691 20100609 ru
MedvedevRussiaE 914990 20100609 en
KremlinRussia 2139735 20100623 ru
PUTIN_VLADIMIR 59264 20100820 en
KremlinRussia_E 296598 20101021 en
Putin_V_V 17753 20110315 en
ElHijoDePutin 17076 20110720 es
Prote 24517 20110823 ru
PutinRF 1152865 20111216 ru
vvp_kreml 438248 20120106 ru
putin_off_ 10306 20120322 ru
PutinRF_Eng 227715 20121107 en
DarthPutinKGB 25589 20121119 en
Identity Use and Misuse of Public Persona on Twitter
169

takes to an open site http://putinism_net. The latter is
run in South- America and offers contents critical to
Putin.
Commentary profiles are used to discuss or state
opinions about the current events with a
concentration on Vladimir Putin or Russian politics.
25 commentary profiles were found.
Fan profiles are built for expressing admiration,
respect etc. towards Putin. Six fan profiles were
identified.
Parody profiles are set up to humorously
counterfeit other people, characters, groups or
objects (Highfield, 2015). Highfield (2015) classifies
parody profile tweets into five groups and states that
they not only post character-specific tweets but also
mention current or newsworthy subjects, trending
topics (i.e. popular hashtags), or post sponsored or
self-promotional comments framed in the context of
the fictional universe or the character’s stereotype.
They may reach more than million followers (i.e.
@Lord_Voldemort7). In our data set, 58 parody
profiles were found. Conspicuously, @Plaid_Putin,
@huyloputin and @PutinsEconomy were the
profiles with a significant number of followers,
6582, 6310 and 4458 respectively. @putinbust,
@VovochkaPutin, and @WhoisMistaPutin were the
other popular profiles with 1729, 1108, and 1798
followers respectively.
Six profiles (@putiin_vovka, @putin_ball,
@Putin_bot, @putinkremline, @Vladi_Putin_bot,
@vseh_pereigral) were classified as spam/bots,
since they seemed to post automated tweets. Two of
these profiles - @Putin_bot and @Vladi_Putin_bot -
were stated to be bots in their profile bios.
Campaign/protest profiles are usually built to
voice people or groups who have similar thoughts or
attitudes towards certain events. These may be
elections or changes of legislation. There are over 50
profiles that were established during 2011 and
stopped tweeting before July 2012. @Putin_Rus
only tweeted on Nov. 5, 2011 four tweets, but
gathered over 2000 followers. Profiles like
@PinkestPutin, @putinarainbow stand against the
gay laws in Russia and exemplify the latter category.
There is a detailed analysis in (Spaiser et al., 2014)
of the latest election campaigns in Russia in 2011
and 2012 and the role of Twitter in them.
Backchannel’s are the Twitter profiles for public
journals like radio shows, books, where they get in
touch with their audience. In our data set there were
7 backchannel profiles which were set for the books,
websites or documentaries concerning Vladimir
Putin (e.g. @MrPutinBook, @i_putin,
@PutinsKiss).
There were 41 profiles that would publish tweets
against Putin (e.g. @putinvor, @SayNoToMrPutin,
@StopPutinstop). Several profiles with ‘stop’ and
‘putin’ in the screen name were created in 2014-
2015 that clearly are pro-Ukrainian and comment the
crisis from the Ukrainian perspective.
There were 79 impersonation profiles, and they
showed a different nature. Some profiles’ tweets
were a mixture of personal posts, and posts from
‘Putin’s mouth’, and news about Putin.
@ComradePutin, @UncleVladimir, and
@VladPutin2013 are examples of these kinds of
profiles.
Some users constructed a modified image of
Vladimir Putin reflecting on how they perceived
him. The impersonations accentuated different
personality characteristics, or public presences of
Vladimir Putin. These included, but were not limited
to a swanky, drunkard, conceited, athletic, sexy,
alpha dog, or gay character. They voiced subtle
tease, or disapproval of his politics or his public
image in their messages. Yet most of them did this
in a humorous way, in the manner of a caricature.
There were also 2 accounts established for
school project and to do research: @PutinStat and
@PutinAP.
Figure 1: Profile Category Distribution for Putin related
profiles.
4.2 Network Analysis in the Putin
Related Set (Spring 2015
Collection)
Based on the follower data we calculated the
followee relation inside our data set. That is, we
found out which profile follows another profile at
least one way inside the data set.
Figure 2 shows mutual followers graph. There
are 57 nodes, and 107 edges. The diameter of the
graph equals to 4. Graph consists of 15 components.
The largest one consists of 21 nodes and 54 edges,
and then there are two components of 6-nodes, two
three-nodes components, and 10 two-node
components. Component that consists of 6 nodes is
WEBIST 2016 - 12th International Conference on Web Information Systems and Technologies
170

formed from four verified profiles, belonging to the
Russian government: @MedvedevRussia,
@MedvedevRussiaE, @KremlinRussia, and
@KremlinRussia_E, then there is verified profile of
@BarackObama, and @Putin, that is not verified.
Figure 3 shows the degree distribution of the graph.
@putinizer has the highest degree equal to 10, i.e. it
is mutually following 10 other profiles. The average
degree is 1,864.
There are 227 profiles in the data set that follow
at least one other profile in the same set. The graph
induced from 582 profiles consists of 227 nodes and
829 edges. It is depicted in Figure 4. Maximal
indegree equals to 98 (belongs to
@MedvedevRussia), maximal outdegree equals to
92 (belonging to @putin_vovchik). This means that
98 profiles in the data set follow @MedvedevRussia
and @putin_vovchick follows 92 other profiles in
the data set.
Figure 2: Mutual followers graph.
Figure 3: Degree distribution of the mutual followers.
Figure 4: Induced graph of the follower relationships in
the Putin related data set.
4.3 Tweet Analysis in the Putin Related
Set (Spring Collection)
We have collected user timelines (message streams)
for the users mentioned above. In total, timelines for
541 users were collected, but for the rest we could
not collect tweets. Overall, 593411 tweets were
collected. Of those, 127294 were retweets of an
earlier tweet. Some of the tweets contain
coordinates, attached to them in the ‘geo’ field.
There were 7194 such tweets, sent by 48 different
profiles.
Table 3 presents countries, extracted from these
tweets, from which the (re)tweets were sent. The
countries were matched against coordinates using
reverse geocoding.
Table 3: Countries presented in the data set.
0
50
123456789
Frequency
Degree
Country Count Country Count
USA 2979 Nigeria 17
Russia 1983 Spain 12
Ukraine 1154 Germany 11
South Africa 304 Austria 8
Philippines 140 Mexico 6
Brazil 119 Ghana 4
France 105 Vietnam 1
Canada 97 Belgium 1
China 78 South Korea 1
Belarus 37 UK 1
Netherlands 32 Kazakhstan 1
Italy 26 Myanmar 1
Identity Use and Misuse of Public Persona on Twitter
171

Figure 5 shows these latitude/longitude points
plotted on the map. 42 profiles have tweeted all geo-
coded tweets from one country, six profiles have
posted tweets from different countries.
@MedvedevRussia has posted tweets from
Kazakhstan, Myanmar, Belarus, Russia, Vietnam,
South Korea, and China. Two users posted tweets
from 3 countries, and three profiles posted tweets
from two different countries.
Figure 5: Coordinates of tweets on the map.
4.4 Profile and Tweet Types in the
Erdogan Related Set (Autumn 2015
Collection)
In total there were 199 profiles related to Recep
Tayyip Erdogan, the Turkish president. Erdogan
related profiles were classified using the Putin
classification as a basis. However, the analysis
resulted in fewer number of profile categories, and
they differed in terms of some characteristics. Figure
6 displays the distribution of the Erdogan related
profiles in this categorization.
Figure 6: Profile Category Distribution for Erdogan
profiles.
As may be seen from Figure 6, no backchannel
or project profiles were found in the data set. The
juxtaposition of the classes according to their profile
frequency is as follows: impersonation, personal, fan
accounts, adversary accounts, parodies, adverts,
campaign/protest accounts, commentaries,
newsfeeds, officials and spam/bots. The allotting of
the accounts among the categories were different
than Putin accounts; and the most common language
was Turkish with 152 profiles, followed by English
with 25 profiles, German with 11 profiles, Dutch
with 5 profiles, Arabic with 4 profiles and French
with 2 profiles.
There are 2 verified profiles representing him:
@rterdogan_ar and @RT_Erdogan. There were 2
newsfeeds (@Erdogan_English and @
tayiperdogann), and 3 commentary profiles
(@RecepTayipE_53, @RajabErdogan_AR,
@TayyipErdoganCB). The 22 fan profiles found in
the profile set also included profiles that were
created to express personal admiration for Erdogan
but did not form a fan community. The 7 parody
profiles were as follows: @RecepErdogan_Ar,
@Tvetci, @RT_ErdoganSpoof, @cbRT_Erdogan1,
@IamRT_ERDOGAN, @Para_erdogan,
@CB_ERDOGAN_. 53 of the 120 impersonation
profiles were empty without any tweets. The four
campaign/protest profiles were established to
promote #dershanemolmasaydı,
#Hepimiz_Takipleselim and #dvltialiosman
hashtags; almost all of their messages were
accompanied by these hashtags, if not they
communicated related messages. 5 advert profiles
were publicizing websites such as
http://www.gamelnet.com/, http://www.nobleandroy
al.com/ and http://www.habera.com/. The two spam
profiles found in the profile set were @jzischke and
@agacili; @agacili profile were increasing the
mentions of @CoolRuhikiziniz profile that is
already suspended. There were 23 personal profiles
without any posts about Recep Tayyip Erdogan; and
9 adversary profiles.
4.5 Number of Namesake Profiles for
the Entire G20 Data Set
The data shows that all G20 leaders have profiles
created (mis)using their digital identity. Table 1
displays the number of Twitter profiles using at least
some part of their digital identity. As may be seen on
the table, Narendra Modi, David Cameron, and
Barack Obama have the highest number of profiles
in descending order. The increase in Putin profiles
from 579 to 786 demonstrates the continuity of this
phenomenon. However, since spring 2015
collection, 67 Putin accounts were closed, 42 of
which were suspended by Twitter from the first
Putin profile set. 34 of the suspended accounts were
personal accounts of type namesake. This points out
that Twitter is following its policy on impersonation,
and keeps accounts that are expressing opinions on a
WEBIST 2016 - 12th International Conference on Web Information Systems and Technologies
172

public person.
Randomly selected 10 percent of the G20
accounts (659 profiles) were grouped taking the
Putin classification as basis. 40 of these profiles
were excluded from categorization as they were
found irrelevant to the subject person. The resulting
classification of this hybrid set is shown in Figure 7.
Figure 7: Profile Category Distribution for Hybrid
Random Profile Set.
5 RELATED WORK
Approaching Twitter’s usage from the angle we
have in this paper is rare. There are, however, many
papers that are relevant in understanding the
‘hundreds of Putin and Obama phenomenon’. One
of the first papers that categorised users is (Java et
al., 2007). The authors identified three major
profile/user categories, information source, friends
and information seeker. Certainly, the most followed
non-verified profiles like @PutinRF are information
sources for the followers. Those profiles that follow
each other (see Fig 3.) could be understood as
‘Friends’. It is obvious that in case of Putin, Erdogan
or Obama there are also political reasons for
establishing profiles. A thorough analysis
concerning the use of Twitter in recent Russian
politics is (Spaiser et al., 2014), although it does not
contain directly the analysis of the ‘Putin profiles’.
In (Bruns and Highfield, 2013) the authors discuss
the use or Twitter in political campaigns in Australia
and in (Peterson, 2012) Peterson discusses the use of
Twitter in US political campaigns. Another category
we found are parody profiles. In a recent paper
(Highfield, 2015) Highfield discusses the parody
profiles in Twitter. This analysis is relevant
especially for those profiles that are tagged as
parody, but also for other profiles that contain jokes
around and about the leaders. A part of the profiles
we found are clearly sexually motivated. A recent
article (Reynolds, 2015) discusses straight men
seeking men and the formation of sexual identity in
virtual space. As concerns wider categorisation of
Twitter profiles, we found (Barash and Kelly, 2012)
that introduces several categories, but based on the
used hashtags in the tweets. Another work (Procter
et al., 2013) categorises profiles in the context of
riots in UK. The authors used circa 10 different
profile categories, like riot profiles, bloggers,
journalists, activists, police profiles, politicians, etc.
They also categorise police profiles into many
subcategories based on whether they are run by a
local police or higher tiers of the police forces.
6 CONCLUSIONS
We have analyzed in this paper Twitter profiles that
carry in their screen name or profile name some
parts of the digital identity of a G20 leader. The first
collection was performed in spring 2015, where we
only collected data related to Vladimir Putin. In
Nov. - Dec. 2015 we collected another data set,
where in addition to Putin we collected profiles
related to other G20 leaders. We found over 6000
profiles from Twitter that qualify. The latter
collection shows that all important leaders get
Twitter accounts that (mis)use their digital identities.
In the spring 2015 collection we included two
verified profiles of the Russian president
@KremlinRussia and @KremlinRussia_E. We
added the two verified profiles of Prime Minister
Medvedev and that of president Obama to see who
follows whom. Thus, we had 5 verified profiles in
the data set and 577 non-verified. We categorized
the profiles according to the information on profile
bios, and the content of the tweets, profile names
and profile pictures. The resulting classification
contained categories official, newsfeed,
commentary, fan profile, parody, impersonation,
campaign/protest, advert, spam/bot, personal,
backchannel, adversary profile and project/research.
Most profiles fell into the category personal or
impersonation.
Looking at the impact of the many profiles based
on the follower numbers it is clear that it is rather
small in the spring 2015 collection. The average
number of followers was 654 among those that have
less than 10000 followers. There are only 13 non-
verified profiles in that data set that have more than
10000 followers. Based on the content of the tweets
one can argue that most of those 13 profiles have a
neutral or positive sentiment towards president
Putin. The two verified Russian profiles controlled
by Kremlin have together 5.76 million followers, out
of which 1.43 million are common. Thus, the
Identity Use and Misuse of Public Persona on Twitter
173

number of distinct followers of Kremlin-controlled
Russian Twitter profiles is over 4 million. The
corresponding verified English profiles have 0.9 and
0.3 million followers, out of which 0.17 million are
common. Thus circa 1 million users follow the
English verified profiles controlled by Kremlin.
Altogether, there are 4.84 million distinct users who
follow one or more of the four verified, Kremlin-
controlled Twitter profiles. Thus, compared to the
other profiles in the data set, these four profiles have
the strongest influence, if we measure this by the
number of followers in our data set.
Apart from the follower analysis, we created a
tentative categorisation of the Putin related profiles.
To the best of our knowledge this is the first attempt
in this direction, i.e. analysing Twitter profiles
around one famous person. We arrived at twelve
tentative categories. In practice all of them are
known from the earlier research. We did not yet
perform a thorough analysis of the contents of the
collected tweets. It would shed more light especially
to the nature of almost 300 ‘personal’ profiles. This
is an item for the future work.
The autumn 2015 data set shows that Putin has
gathered 200 more related accounts in half a year.
Why and what kind of profiles is for further study.
The tentative categorisation developed for the
spring 2015 data set concerning Putin related
account seems to be valid for the Erdogan related
profile set as well.
As concerns the theories that might explain the
phenomenon, there seems to be no one theory that
would explain the ‘hundreds of Putins and Obamas
phenomenon’ on Twitter. Some theories explain a
part of the profiles and the activity on them; some
profiles are clearly politically motivated, some are
parody profiles, some fan accounts etc. The online
disinhibition effect and the anonymity’s enabling
power for free speech were seen in many of the
parody, commentary, adversary, impersonation and
campaign/protest accounts. In addition advert and
spam/bot accounts were examples of malicious
anonymity usage. Over 50 profiles were created
during the elections in Russia 2011-2012 and 74
became silent before mid-2012. Thus, they are most
probably campaign profiles. Most geo-coded tweets
came from USA in the spring 2015 data set, but this
cannot be generalized to the entire data set, because
only one per cent of the collected tweets had the
coordinates in ‘geo’ attribute. Many profiles had
location in the profile, but we did not yet match this
with the tweets or their content. The languages used
on the profiles were mostly Russian or English in the
spring 2015 data set. In the autumn 2015 data set it
varied more, because we had also Saudi-Arabia,
Japan and Korea included into the data set.
The study was useful in terms of improving the
understanding of social media culture, and usage of
public identities on online social networks. In the
future we will delve deeper into the autumn 2015
data and will also analyse the follower set relations
for the famous persons.
ACKNOWLEDGEMENTS
The work of the two first authors was supported in
part by the Academy of Finland, grant #268078
(MineSocMed).
REFERENCES
Akdeniz, Y., 2002. Anonymity, Democracy, and
Cyberspace. Soc. Res. 69, 223–237.
Altheide, D.L., 2000. Identity and the Definition of the
Situation in a Mass-Mediated Context. Symb. Interact.
23, 1–27. doi:10.1525/si.2000.23.1.1.
Ardia, D.S., 2012. Reputation in a Networked World:
Revisiting the Social Foundations of Defamation Law
(SSRN Scholarly Paper No. ID 1689865). Social
Science Research Network, Rochester, NY.
Barash, V., Kelly, J., 2012. Salience vs. Commitment:
Dynamics of Political Hashtags in Russian Twitter
(SSRN Scholarly Paper No. ID 2034506). Social
Science Research Network, Rochester, NY.
Belk, R.W., 2013. Extended Self in a Digital World. J.
Consum. Res. 40, 477–500. doi:10.1086/671052.
Bodle, R., 2013. The ethics of online anonymity or
Zuckerberg vs. “Moot.” ACM SIGCAS Comput. Soc.
43, 22–35. doi:10.1145/2505414.2505417.
Bruns, A., Highfield, T., 2013. Political Networks on
Twitter. Inf. Commun. Soc. 16, 667–691.
doi:10.1080/1369118X.2013.782328.
Castells, M., 2007. Communication, Power and Counter-
power in the Network Society. Int. J. Commun. 1, 29.
China, I., 2015. Baidu Post Bar More MAUs Than Weibo
in Jan 2015 [WWW Document]. China Internet
Watch. URL http://www.chinainternetwatch.com/1262
8/baidu-post-bar-more-maus-than-weibo-jan-2015/
(accessed 10.13.15).
Christopherson, K.M., 2007. The positive and negative
implications of anonymity in Internet social
interactions: “On the Internet, Nobody Knows You’re
a Dog.” Comput. Hum. Behav., Including the Special
Issue: Education and Pedagogy with Learning Objects
and Learning Designs 23, 3038–3056.
doi:10.1016/j.chb.2006.09.001.
Fredheim, R., Moore, A., Naughton, J., 2015. Anonymity
and Online Commenting: An Empirical Study. SSRN
Electron. J. doi:10.2139/ssrn.2591299.
WEBIST 2016 - 12th International Conference on Web Information Systems and Technologies
174

Goffman, E., 1959. The presentation of self in everyday
life 1–17.
Hearn, A., 2008. `Meat, Mask, Burden`: Probing the
contours of the branded `self`. J. Consum. Cult. 8,
197–217. doi:10.1177/1469540508090086.
Highfield, T., 2015. News via Voldemort: Parody
accounts in topical discussions on Twitter. New Media
Soc. 1–18. doi:10.1177/1461444815576703.
Hollenbeck, C.R., Zinkhan, G.M., 2006. Consumer
Activism on the Internet: The Role of Anti-brand
Communities. Adv. Consum. Res. 33, 479–485.
Java, A., Song, X., Finin, T., Tseng, B., 2007. Why we
twitter: understanding microblogging usage and
communities. ACM Press, pp. 56–65.
doi:10.1145/1348549.1348556.
Kaye, D., 2015. Report of the Special Rapporteur on the
promotion and protection of the right to freedom of
opinion and expression. Human Rights Council.
Kiss, J., 2008. How Obama’s online campaign helped win
him the presidency [WWW Document]. The Guardian.
URL http://www.theguardian.com/media/2008/nov/10
/obama-online-strategy (accessed 12.4.15).
Kling, R., Lee, Y.-C., Teich, A., Frankel, M.S., 1999.
Assessing Anonymous Communication on the
Internet: Policy Deliberations. Inf. Soc. 15, 79–90.
doi:10.1080/019722499128547.
Lapidot-Lefler, N., Barak, A., 2012. Effects of anonymity,
invisibility, and lack of eye-contact on toxic online
disinhibition. Comput. Hum. Behav. 28, 434–443.
doi:10.1016/j.chb.2011.10.014.
Leary, M., Kowalski, R., 1990. Impression management:
A literature review and two-component model.
Psychol. Bull. 107, 34–47. doi:10.1037/0033-
2909.107.1.34.
Marshall, P.D., 2010. The promotion and presentation of
the self: celebrity as marker of presentational media.
Celebr. Stud. 1, 35–48.
doi:10.1080/19392390903519057.
Miller, C.C., 2008. How Obama’s Internet Campaign
Changed Politics - The New York Times [WWW
Document]. N. Y. Times. URL
http://bits.blogs.nytimes.com/2008/11/07/how-obamas
-internet-campaign-changed-politics/?_r=1 (accessed
12.4.15).
Miller, H., 1995. The Presentation of Self in Electronic
Life: Goffman on the Internet, in: Embodied
Knowledge and Virtual Space Conference Goldsmiths’
College. Goldsmiths’ College, London, UK.
Peterson, R.D., 2012. To tweet or not to tweet: Exploring
the determinants of early adoption of Twitter by House
members in the 111th Congress. Soc. Sci. J., Special
Issue: National and state politics: A current
assessment 49, 430–438.
doi:10.1016/j.soscij.2012.07.002.
Procter, R., Crump, J., Karstedt, S., Voss, A., Cantijoch,
M., 2013. Reading the riots: what were the police
doing on Twitter? Polic. Soc. 23, 413–436.
doi:10.1080/10439463.2013.780223.
Reynolds, C., 2015. “I Am Super Straight and I Prefer
You be Too” Constructions of Heterosexual
Masculinity in Online Personal Ads for “Straight”
Men Seeking Sex With Men. J. Commun. Inq. 1–19.
doi:10.1177/0196859915575736.
Santana, A.D., 2014. Virtuous or Vitriolic: The effect of
anonymity on civility in online newspaper reader
comment boards. Journal. Pract. 8, 18–33.
doi:10.1080/17512786.2013.813194.
Scott, C.R., 2004. Benefits and Drawbacks of Anonymous
Online Communication: Legal Challenges and
Communicative Recommendations. Free Speech
Yearb. 41, 127–141.
doi:10.1080/08997225.2004.10556309.
Skogerbø, E., Krumsvik, A.H., 2015. Newspapers,
Facebook and Twitter: Intermedial agenda setting in
local election campaigns. Journal. Pract. 9, 350–366.
doi:10.1080/17512786.2014.950471.
Spaiser, V., Chadefaux, T., Donnay, K., Russmann, F.,
Helbing, D., 2014. Social Media and Regime Change:
The Strategic Use of Twitter in the 2011–12 Russian
Protests (SSRN Scholarly Paper No. ID 2528102).
Social Science Research Network, Rochester, NY.
Suler, J., 2004. The Online Disinhibition Effect.
Cyberpsychol. Behav. 7, 321–326.
doi:10.1089/1094931041291295.
Suler, J.R., 2002. Identity Management in Cyberspace. J.
Appl. Psychoanal. Stud. 4, 455–459.
doi:10.1023/A:1020392231924.
Sundén, J., 2003. Material virtualities: approaching online
textual embodiment, Digital formations. P. Lang, New
York.
Twitter Help Center, 2014. Impersonation Policy.
Twitter, Inc., 2005. Company | About [WWW Document].
Twitter About. URL https://about.twitter.com/comp
any (accessed 10.13.15).
Zhao, S., Grasmuck, S., Martin, J., 2008. Identity
construction on Facebook: Digital empowerment in
anchored relationships. Comput. Hum. Behav. 24,
1816–1836. doi:10.1016/j.chb.2008.02.012.
Identity Use and Misuse of Public Persona on Twitter
175