Sentiment of Emojis

Abstract

There is a new generation of emoticons, called emojis, increasingly used in
mobile communications and social media. In the last two years, over ten billion
of emojis were used on Twitter. Emojis are Unicode graphic symbols, used as a
shorthand to express concepts and ideas. In contrast to a small number of
well-known emoticons which carry clear emotional contents, there are hundreds
of emojis. But what is their emotional contents? We provide the first emoji
sentiment lexicon, called Emoji Sentiment Ranking, and draw a sentiment map of
the 751 most frequently used emojis. The sentiment of emojis is computed from
the sentiment of tweets in which they occur. We have engaged 83 human
annotators to label over 1.6 million tweets in 13 European languages by the
sentiment polarity (negative, neutral, or positive). About 4\% of the annotated
tweets contain emojis. The sentiment analysis of emojis yields several
interesting conclusions. It turns out that most of the emojis are positive,
especially the most popular ones. The sentiment distribution of the tweets with
and without emojis is significantly different. The inter-annotator agreement on
the tweets with emojis is higher. Emojis tend to occur at the end of the
tweets, and their sentiment polarity increases with the distance. We observe no
significant differences in emoji rankings between the 13 languages, and propose
our Emoji Sentiment Ranking as a European language-independent resource for
automated sentiment analysis. Finally, the paper provides a formalization of
sentiment and novel visualization in the form of a sentiment bar.

Full text

arXiv:1509.07761v2 [cs.CL] 8 Dec 2015
Sentiment of Emojis
Petra Kralj Novak, Jasmina Smailovi´c, Borut Sluban, Igor Mozetiˇc
Joˇzef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia
* Petra.Kralj.Novak@ijs.si, Igor.Mozetic@ijs.si
Abstract
There is a new generation of emoticons, called emojis, that is increasingly being used in mobile
communications and social media. In the past two years, over ten billion emojis were used on Twitter.
Emojis are Unicode graphic symbols, used as a shorthand to express concepts and ideas. In contrast
to the small number of well-known emoticons that carry clear emotional contents, there are hundreds
of emojis. But what are their emotional contents? We provide the first emoji sentiment lexicon, called
the Emoji Sentiment Ranking, and draw a sentiment map of the 751 most frequently used emojis. The
sentiment of the emojis is computed from the sentiment of the tweets in which they occur. We engaged
83 human annotators to label over 1.6 million tweets in 13 European languages by the sentiment
polarity (negative, neutral, or positive). About 4% of the annotated tweets contain emojis. The
sentiment analysis of the emojis allows us to draw several interesting conclusions. It turns out that
most of the emojis are positive, especially the most popular ones. The sentiment distribution of the
tweets with and without emojis is significantly different. The inter-annotator agreement on the tweets
with emojis is higher. Emojis tend to occur at the end of the tweets, and their sentiment polarity
increases with the distance. We observe no significant differences in the emoji rankings between the 13
languages and the Emoji Sentiment Ranking. Consequently, we propose our Emoji Sentiment Ranking
as a European language-independent resource for automated sentiment analysis. Finally, the paper
provides a formalization of sentiment and a novel visualization in the form of a sentiment bar.
1 Introduction
An emoticon, such as ;-), is shorthand for a facial expression. It allows the author to express her/his
feelings, moods and emotions, and augments a written message with non-verbal elements. It helps to
draw the reader’s attention, and enhances and improves the understanding of the message. An emoji
is a step further, developed with modern communication technologies that facilitate more expressive
messages. An emo ji is a graphic symbol, ideogram, that represents not only facial expressions, but also
concepts and ideas, such as celebration, weather, vehicles and buildings, food and drink, animals and
plants, or emotions, feelings, and activities.
Emojis on smartphones, in chat, and email applications have become extremely popular worldwide.
For example, Instagram, an online mobile photo-sharing, video-sharing and social-networking
platform, reported in March 2015 that nearly half of the texts on Instagram contained emo jis [1]. The
use of emojis on the SwiftKey Android and iOS keybords, for devices such as smartphones and tablets,
was analyzed in the SwiftKey Emoji Report [2], where a great variety in the popularity of individual
emojis, and even between countries, was reported. However, to the best of our knowledge, no
large-scale analysis of the emotional content of emojis has been conducted so far.
Sentiment analysis is the field of study that analyzes people’s opinions, sentiments, evaluations,
attitudes, and emotions from a text [3,4]. In analyzing short informal texts, such as tweets, blogs or
comments, it turns out that the emoticons provide a crucial piece of information [5–12]. However,
1

emojis have not been exploited so far, and no resource with emoji sentiment information has been
provided.
In this paper we present the Emoji Sentiment Ranking, the first emoji sentiment lexicon of 751
emojis. The lexicon was constructed from over 1.6 million tweets in 13 European languages, annotated
for sentiment by human annotators. In the corpus, probably the largest set of manually annotated
tweets, 4% of the tweets contained emojis. The sentiment of the emojis was computed from the
sentiment of the tweets in which they occur, and reflects the actual use of emojis in a context.
Background. An emoticon is a short sequence of characters, typically punctuation symbols. The
use of emoticons can be traced back to the 19th century, when they were used in casual and humorous
writing. The first use of emoticons in the digital era is attributed to professor Scott Fahlman, in a
message on the computer-science message board of Carnegie Mellon University, on September 19, 1982.
In his message, Fahlman proposed to use :-) and :-( to distinguish jokes from more serious posts.
Within a few months, the use of emoticons had spread, and the set of emoticons was extended with
hugs and kisses, by using characters found on a typical keyboard. A decade later, emoticons had found
their way into everyday digital communications and have now become a paralanguage of the web [6].
The word ‘emoji’ literally means ‘picture character’ in Japanese. Emojis emerged in Japan at the
end of the 20th century to facilitate digital communication. A number of Japanese carriers (Softbank,
KDDI, DoCoMo) provided their own implementations, with incompatible encoding schemes. Emojis
were first standardized in Unicode 6.0 [13]—the core emoji set consisted of 722 characters. However,
Apple’s support for emojis on the iPhone, in 2010, led to global popularity. An additional set of about
250 emojis was included in Unicode 7.0 [14] in 2014. As of August 2015, Unicode 8.0 [15] defines a list
of 1281 single- or double-character emoji symbols.
Related work. Sentiment analysis, or opinion mining, is the computational study of people’s
opinions, sentiments, emotions, and attitudes. It is one of the most active research areas in
natural-language processing and is also extensively studied in data mining, web mining, and text
mining [3, 4]. The growing importance of se ntiment analysis coincides with the growth of social media,
such as Twitter, Facebook, book reviews, forum discussions, blogs, etc.
The basis of many sentiment-analysis approaches is the sentiment lexicons, with the words and
phrases classified as conveying positive or negative sentiments. Several general-purpose lexicons of
subjectivity and sentiment have been constructed. Most sentiment-analysis research focuses on English
text and, consequently, most of the resources developed (such as sentiment lexicons and corpora) are
in English. One such lexical resource, explicitly devised to support sentiment classification and opinion
mining, is SentiWordNet 3.0 [16]. SentiWordNet extends the well-known WordNet [17] by associating
each synset with three numerical scores, describing how ‘objective’, ‘positive’, and ‘negative’ the terms
in the synset are.
Emoticons have proved crucial in the automated sentiment classification of informal texts [5–12]. In
an early work [10], a basic distinction between positive and negative emoticons was used to
automatically generate positive and negative samples of texts. These samples were then used to train
and test sentiment-classification models using machine learning techniques. The early results
suggested that the sentiment conveyed by emoticons is both domain and topic independent. In later
work, these findings were applied to automatically construct sets of positive and negative
tweets [8, 18, 19], and sets of tweets with alternative sentiment categories, such as the angry and sad
emotional states [11]. Such emoticon-labeled sets are then used to automatically train the sentiment
classifiers. Emoticons can also be exploited to extend the more common features used in text mining,
such as sentiment-carrying words. A small set of emoticons has already been used as additional
features for polarity classification [8, 20]. A sentiment-analysis framework that takes explicitly into
account the information conveyed by emoticons is proposed in [6].
There is also research that analyzes graphical emoticons and their sentiment, or employs them in a
sentiment classification task. The authors in [21] manually mapped the emoticons from Unicode 8.0 to
nine emotional categories and performed the sentiment classification of tweets, using both emoticons
2

and bag-of-words as features. Ganesan et al. [22] presents a system for adding the graphical emoticons
to text as an illustration of the written emotions.
Several studies have analyzed emotional contagion through posts on Facebook and showed that the
emotions in the posts of online friends influence the emotions expressed in newly generated
content [23–26]. Gruzd et al. [27] examined the spreading of emotional content on Twitter and found
that the positive posts are retweeted more often than the negative ones. It would be interesting to
examine how the presence of emojis in tweets affects the spread of emotions on Twitter, i.e., to relate
our study to the field of emotional contagion [28].
Contributions. Emojis, a new generation of emoticons, are increasingly being used in social media.
Tweets, blogs and comments are analyzed to estimate the emotional attitude of a large fraction of the
population to various issues. An emo ji sentiment lexicon, provided as a result of this study, is a
valuable resource for automated sentiment analysis. The Emoji Sentiment Ranking has a format
similar to SentiWordNet [16], a publicly available resource for opinion mining, used in more than 700
applications and studies so far, according to Google Scholar. In addition to a public resource, the
paper provides an in-depth analysis of several aspects of emo ji sentiment. We draw a sentiment map of
the 751 emojis, compare the differences between the tweets with and without emojis, the differences
between the more and less frequent emojis, their positions in tweets, and the differences between their
use in the 13 languages. Finally, a formalization of sentiment and a novel visualization in the form of a
sentiment bar are presented.
2 Results and Discussion
2.1 Emoji sentiment lexicon
The sentiment of emojis is computed from the sentiment of tweets. A large pool of tweets, in 13
European languages, was labeled for sentiment by 83 native speakers. Sentiment labels can take one of
three ordered values: negative ≺neutral ≺positive. A sentiment label, c, is formally a discrete,
3-valued variable c∈ {−1,0,+1}. An emo ji is assigned a sentiment from all the tweets in which it
occurs. First, for each emoji, we form a discrete probability distribution (p−,p0,p+). The sentiment
score sof the emoji is then computed as the mean of the distribution. The components of the
distribution, i.e., p−,p0, and p+denote the negativity, neutrality, and positivity of the emoji,
respectively. The probability pcis estimated from the number of occurrences, N, of the emoji in
tweets with the label c. Note that an emoji can occur multiple times in a single tweet, and we count
all the occurrences. A more detailed formalization of the sentiment representation can be found in the
Methods section.
We thus form a sentiment lexicon of the 751 most frequent emojis, called the Emo ji Sentiment
Ranking. The complete Emoji Sentiment Ranking is available as a web page at
http://kt.ijs.si/data/Emoji_sentiment_ranking/. The 10 most frequently used emojis from the
lexicon are shown in Fig 1.
First we address the question of whether the emojis in our lexicon are representative. We checked
Emojitracker (http://emojitracker.com/), a website that monitors the use of emojis on Twitter
in realtime. In the past two years, Emojitracker has detected almost 10 billion emojis on Twitter!
From the ratio of the number of emoji occurrences and tweets in our dataset (∼2.3), we estimate that
there were about 4 billion tweets with emojis. In our dataset of about 70,000 tweets, we found 969
different emojis, 721 of them in common with Emojitracker.
We compared the emojis in both sets, ordered by the number of occurrences, using Pearson’s [29]
and Spearman’s rank [30] correlation. We successively shorten our list of emo jis by cutting off the
least-frequent emojis. The results for two thresholds, N≥1 and 5, with the highest correlation
coefficients, are shown in Table 1. Both correlation coefficients are high, significant at the 1% level,
thus confirming that our list of emojis is indeed representative of their general use on Twitter.
Between the two options, we decided to select the list of emojis with at least 5 occurrences, resulting
3

h#H2RX1KQDBb2MiBK2Mi`MFBM;X1KQDBLSQbBiBQMp−p0p+sLK2قR9-ekkyX3yyXk8yXkNyX9dyXkk7+2rBi?i2`bQ7DQvƪ3-y8yyXd9yXy9yXRdyXdNyXd8?2pv#H+F?2`iłd-R99yXd8yXy9yXkdyXeNyXee#H+F?2`ibmBi٨e-j8NyXdeyXy8yXkkyXdjyXe3bKBHBM;7+2rBi??2`i@b?T2/2v2bڷ8-8keyX3yyX99yXkkyXj9@yXyNHQm/Hv+`vBM;7+2ڄj-e93yX38yXy8yXRNyXd8yXdy7+2i?`QrBM;FBbbٝj-R3eyX3RyXyeyXk9yXdyyXe9bKBHBM;7+2rBi?bKBHBM;2v2bэk-Nk8yX3yyXyNyXk8yXeeyX8eQF?M/bB;MӲk-9yyyXdeyXy9yXkNyXedyXejirQ?2`ibѐk-jjeyXd3yXRyyXkdyXekyX8k+HTTBM;?M/bbB;Ma2MiBK2Mi/Bbi`B#miBQMdbB;MB}+MiHv/Bz2`2Mi#2ir22Mir22ibrBi?M/rBi?Qmi2KQDBb3h#H2kNh#H2kXh?2MmK#2`M/T2`+2Mi;2Q7M2;iBp2-M2mi`H-M/TQbBiBp2hrBii2`K2bb;2brBi?M/rBi?Qmi2KQDBbX1KQDB/ib2iLQ@2KQDB/ib2iL2;iBp2Rk-R8eURd-98WV9Ry-jyRUke-ydWVL2mi`HRN-Nj3Uk3-ekWV83d-jjdUjd-jRWVSQbBiBp2jd-8dNU8j-N9WV8de-9k9Uje-ekWVa2MiBK2MiK2MhQiHeN-edjR-8d9-yekAMi2`@MMQiiQ`;`22K2MiRybB;MB}+MiHv/Bz2`2Mi#2ir22Mir22ibrBi?M/rBi?Qmi2KQDBbh#H2jRRh#H2jXh?2BMi2`@MMQiiQ`;`22K2MiQMir22ibrBi?M/rBi?Qmi2KQDBbK2bm`2/BMi2`KbQ7i?`222pHmiBQMK2i`B+bXhr22ibrBi?2KQDBbhr22ibrBi?Qmi2KQDBb++m`+vUVUV++m`+vUV*Q?2MǶbr2B;?i2/FTTa2MiBK2MiM/TQbBiBQMRk6B;m`2kRja2MiBK2Mi#vHM;m;2R9h#H29R8
SGPakf8
Figure 1. Top 10 emojis. Emojis are ordered by the number of occurrences N. The average
position ranges from 0 (the beginning of the tweets) to 1 (the end of the tweets). pc,c∈ {−1,0,+1},
are the negativity, neutrality, and positivity, respectively. sis the sentiment score.
in the lexicon of 751 emojis. The sentiment scores for the emojis with fewer then 5 occurrences are not
very reliable.
Table 1. Overlap with Emojitracker. Correlations are between the occurrences of emojis in the
Emoji Sentiment Ranking and Emojitracker, for two minimum occurrence thresholds. The numbers in
parenthesis are the emojis that are common in both sets. The correlation values, significant at the 1%
level, are indicated by *.
Tweets Different Pearson Spearman rank
with emojis emo jis used correlation correlation
Emojitracker ∼4 billion 845 / /
Emoji Sent. Rank.
N≥1 69,673 969 (721) 0.945* 0.897*
Emoji Sent. Rank.
N≥5 69,546 751 (608) 0.944* 0.898*
4

2.2 Emoji sentiment map
Before we analyze the properties of the tweets with emojis, we first discuss two visualizations of the
Emoji Sentiment Ranking. Fig 2 shows the overall map of the 751 emojis. The position of an emoji is
determined by its sentiment score sand its neutrality p0. The sentiment score sis in the range
(−1,+1) and is computed as p+−p−. The more positive emo jis are on the right-hand side of the map
(green), while the negative ones are on the left-hand side (red). The neutral emo jis (yellow) span a
whole band around s= 0. The emojis are prevailingly positive, the mean sentiment score is +0.3 (see
Fig 4). The bubble sizes are proportional to the number of occurrences.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Neutrality
Sentiment score
A
B
C
Figure 2. Sentiment map of the 751 emo jis. Left: negative (red), right: positive (green), top:
neutral (yellow). Bubble size is proportional to log10 of the emoji occurrences in the Emoji Sentiment
Ranking. Sections A, B, and C are references to the zoomed-in panels in Fig 3.
A more detailed view of some actual emojis on the map is shown in Fig 3. The most frequent
negative emojis (panel A) are sad faces. On the other hand, the most frequent positive emojis (panel
C) are not only happy faces, but also hearts, party symbols, a wrapped present, and a trophy. Even
more interesting are the neutral emojis (panel B). All of them have a sentiment score around 0, but
the neutrality p0ranges between 0 and 1. The bottom two, with low p0(face with cold sweat, crying
face), are bipolar, with a high negativity and positivity, where p−≈p+. The middle two (flushed face,
bomb) have a uniform sentiment distribution, where p−≈p0≈p+. The top ones, with high p0, are
neutral indeed, symbolized by the yin yang symbol at the very top.
5

0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-0.05 0
0.05
Neutrality
Sentiment score
☯
█
░
0.1
0.15
0.2
0.25
0.3
-0.4 -0.35
-0.3
Neutrality
Sentiment score
0.1
0.15
0.2
0.25
0.3
0.7 0.75 0.8
Neutrality
Sentiment score
A B
C
Figure 3. Emo jis in sections A, B, and C of Fig 2. Shown are emojis that occur at least 100
times in the Emoji Sentiment Ranking. Panel A: negative emojis, panel B: neutral (top) and bipolar
(bottom) emojis, panel C: positive emojis.
2.3 Tweets with and without emojis
In this subsection we analyze the interplay of the human perception of tweets that are with and
without emojis. If we consider the sentiment of a tweet as a rough approximation of its emotional
content, we can ask two questions. Are the tweets with emojis more emotionally loaded? Does the
presence of emojis in tweets have an impact on the human emotional perception of the tweets? We do
not draw any causal conclusions, but report the results of two experiments that indicate that the
answer to both questions is positive.
First, we compare all the manually labeled tweets that are with and without emojis. From the
distribution of the negative, neutral, and positive tweets in both sets, we compute the mean, standard
deviation (sd), and standard error of the mean (sem). The results are shown in Table 2.
We test the null hypothesis that the two populations have equal means. We apply Welch’s
t-test [31] for two samples with unequal variances and sizes. We are aware that the two populations
might not be normally distributed, but Welch’s t-test is robust for skewed distributions, and even
more so for large sample sizes [32]. With t= 87, the degrees of freedom ≫100 (due to large sample
sizes), and the p-value ≈0, the null hypothesis can be rejected. We can conclude, with high
confidence, that the tweets with and without emojis have significantly different sentiment means.
Additionally, the tweets with emojis are significantly more positive (mean = +0.365) than the tweets
without emo jis (mean = +0.106).
Next, we compare the agreement of the human annotators on the tweets with and without emojis.
The Twitter sentiment classification is not an easy task and humans often disagree on the sentiment
labels of controversial tweets. During the process of annotating the 1.6 million tweets, we found that
6

Table 2. Sentiment of tweets with and without emojis. For each set, the mean, sd and sem
are computed from the distribution of negative, neutral, and positive tweets.
Tweets Tweets
Sentiment with emojis without emojis
Negative 12,156 (17.5%) 410,301 (26.1%)
Neutral 19,938 (28.6%) 587,337 (37.3%)
Positive 37,579 (53.9%) 576,424 (36.6%)
Total 69,673 1,574,062
Mean +0.365 +0.106
sd,sem 0.762, 0.0029 0.785, 0.0006
even individual annotators are not consistent with themselves. Therefore, we systematically distributed
a fraction of the tweets to be annotated twice in order to estimate the level of agreement. This
annotator self-agreement is a good indicator of the reliability of the annotator. The inter-annotator
agreement, on the other hand, indicates the difficulty of the task. In the case of emojis, our goal is to
verify whether their presence in tweets correlates with a higher inter-annotator agreement.
There are a number of measures to estimate the inter-annotator agreement. We apply three of
them from two different fields, to ensure robust estimates. The first one, Krippendorff’s
Alpha-reliability [33], generalizes several specialized agreement measures. When the annotators are in
perfect agreement, Alpha = 1, and when the level of agreement equals the agreement by chance,
Alpha = 0. We applied an instance of Alpha that takes into account the ordering of labels and assigns
a higher penalty to more extreme disagreements. For example, a disagreement between the negative
and the positive sentiment is four times as costly as that between the neutral and positive.
The simplest measure of agreement is the joint probability of agreement, also known as Accuracy ,
when evaluating classification models. Accuracy is the number of equally labeled tweets by different
annotators, divided by the total number of tweets labeled twice. It assumes the data labels are
unordered (nominal) and does not take into account the agreement by chance, but it is easy to
interpret.
The third measure comes from the field of machine learning. It is used to evaluate the performance
of classification models against a test set, where the true sentiment label is known. The measure,
F1(−,+), is a standard measure of performance, specifically designed for a 3-valued sentiment
classification [12], where the negative (−) and positive (+) sentiments are considered more important
than the neutral one. Here, we adapt it to estimate the agreement of a pair of annotators.
Table 3. Inter-annotator agreement on tweets with and without emojis. The agreement is
computed in terms of three measures over a subset of tweets that were labeled by two different
annotators.
Agreement Tweets Tweets
measure with emojis without emojis
Alpha 0.597 0.495
Accuracy 0.641 0.583
F1(−,+) 0.698 0.598
No. of tweets
annotated twice 3,547 52,027
Table 3 gives the results of the inter-annotator agreements on the tweets with and without emojis.
Coincidence matrices for both cases are in the Methods section, in Tables 9 and 10, respectively. All
three measures of agreement, Alpha,Accuracy, and F1(−,+), are considerably higher for the tweets
with emojis, by 21%, 10%, and 17%, respectively. We do not give any statistical-significance results,
but it seems safe to conclude that the presence of emo jis has a positive impact on the emotional
perception of the tweets by humans. After all, this is probably the main reason why they are used in
the first place.
7

2.4 Sentiment distribution
In this subsection we analyze the sentiment distribution of the emojis with respect to the frequency of
their use. The question we address is the following: Are the more frequently used emojis more
emotionally loaded? First, in Fig 4 we show the sentiment distribution of the 751 emojis, regardless of
their frequencies. It is evident that the sentiment score of the emojis is approximately normally
distributed, with mean = +0.3, prevailingly positive.
0
10
20
30
40
50
60
70
-1.00
-0.95
-0.90
-0.85
-0.80
-0.75
-0.70
-0.65
-0.60
-0.55
-0.50
-0.45
-0.40
-0.35
-0.30
-0.25
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
Different emojis
Sentiment score
mean
sentiment
score
Figure 4. Distribution of emojis by sentiment score. The mean sentiment score of the 751
emojis (in bins of 0.05) is +0.305.
In Fig 5 we rank the emojis by the number of their occurrences in tweets. The sentiment score of
each emoji is indicated by the color. The zoomed-in section of the first 33 emojis is in Fig 6.
We did not thoroughly analyze the frequency-rank distribution of the emojis. A quick analysis
suggests that the data follows a power law with an exponential cutoff at a rank of about 200. Using a
maximum-likelihood estimator [34], the exponent of the power law is estimated to be −1.3, a relatively
extreme exponent. Even more relevant is the distribution of the emojis on Emojitracker, but this
remains a subject of further research. Here we concentrate on the sentiment distribution.
We define a cumulative distribution function cdf(R) of rank Rover a set of ranked emojis as:
cdf(R) = N(r≤R) = X
r≤R
N(r),
where rdenotes the rank of an emoji, and N(r) the number of occurrences of the emoji at rank r. In
order to compare the higher-ranked emojis (more frequent) with the lower-ranked ones (less frequent),
we define a midpoint rank R1/2, such that:
N(1 ≤r≤R1/2)≈N(R1/2< r ≤751) .
The midpoint rank R1/2partitions the emo jis into two subsets with an approximately equal cumulative
number of occurrences. In the case of the Emoji Sentiment Ranking, the midpoint is at R1/2= 23.
We compute the mean sentiment, sd, and sem of the more frequent and the less frequent emojis.
The results are shown in Table 4.
We test the null hypothesis that the two populations of emojis have equal mean sentiment scores.
Again, we apply Welch’s t-test for two samples with unequal variances, but similar sizes. With
8

1
4
16
64
256
1,024
4,096
16,384
0 100 200 300 400 500 600 700
Occurrences
Emoji rank
Sentiment score
Negative Neutral Positive
R1/2
Figure 5. Distribution of occurrences and sentiment of the 751 emojis. The emojis are
ranked by their occurrence (log scale). The column color indicates the sentiment score. The
partitioning into two equally weighted halfs is indicated by a line at R1/2. The first 33 emo jis are
zoomed-in in Fig 6.
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000

❤
♥
!
"
#
$
%
&
'
☺
♡
)
*
+
✌
,
-
.
/
0
1
2
3
4
5
6
7
8
9
☯
:
Occurrences
Emoji
Negative
Neutral Positive
Sentiment score
Figure 6. Top 33 emojis by occurrence. Column color represents the emoji sentiment score.
t= 100, the degrees of freedom ≫100 (due to large sample sizes), and the p-value ≈0, the null
hypothesis can be rejected. We can conclude, with high confidence, that the more-frequent emojis are
significantly more positive than the less-frequent ones.
This result supports the thesis that the emojis that are used more often are more emotionally
loaded, but we cannot draw any causal conclusion. Are they more positive because they are more
often used in positive tweets, or are they more frequently used, because they are more positive?
9

Table 4. Comparison of the more-frequent with the less-frequent emojis. The emojis (r)
ranked by occurrence N(r) are partitioned into two halves with approximately the same cumulative
number of occurrences.
1st half (r≤23) 2nd half (23 < r) Total
Different emojis 23 728 751
Occurrences (PN(r)) 77,969 78,488 156,457
Sentiment mean +0.463 +0.311 +0.387
sd,sem 0.280, 0.0010 0.319, 0.0011 0.300, 0.0008
2.5 Sentiment and emoji position
Where are the emojis typically placed in tweets? Emoticons such as :-) are used sparsely and typically
at the very end of a sentence. Emojis, on the other hand, appear in groups and not only at the end of
the tweets. Fig 7 shows the average positions of the 751 emojis in the tweets. On average, an emoji is
placed at 2/3 of the length of a tweet.
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Sentiment score
Position
Figure 7. Average positions of the 751 emojis in tweets. Bubble size is proportional to log10 of
the emoji occurrences in the Emoji Sentiment Ranking. Left: the beginning of tweets, right: the end
of tweets, bottom: negative (red), top: positive (green).
Fig 7 also indicates the sentiment of an emoji in relation to its position. In Fig 8 we decompose the
sentiment into its three constituent components and show the regression trendlines.
The linear regression functions in Fig 8 have the following forms:
negativity: p−(d) = 0.20d+ 0.03 (R2= 0.06) ,
neutrality: p0(d) = −0.41d+ 0.66 (R2= 0.14) ,
positivity: p+(d) = 0.21d+ 0.30 (R2= 0.04) ,
where dis the distance from the beginning of the tweets. The functions do not fit the data very well,
but they give some useful insight. At any distance d, and for any subset of emo jis, the component
probabilities add up to 1:
X
c
pc(d) = 1
However, the negativity and positivity increase with the distance, whereas the neutrality decreases.
This means that more emotionally loaded emojis, either negative or positive, tend to occur towards
the end of the tweets.
10

0
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
Negativity
Position
0
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
Neutrality
Position
0
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
Positivity
Position
Figure 8. Negativity, neutrality, and positivity regressed with position (from left to
right). The trendlines are functions pc(d) of the distance dfrom the beginning of the tweets.
2.6 Emojis in different languages
In the final subsection we analyze the use of emojis in the 13 languages processed in this study. Can
the Emoji Sentiment Ranking be considered a universal resource, at least for European languages? Is
the sentiment ranking between the different languages significantly different? The results in Table 5
indicate that the answer to the first question is positive and that there is no evidence of significant
differences between the languages.
Table 5. Emoji sentiment in different languages. The languages are ordered by the number of
different emojis used. Correlations are between the sentiment scores of emojis in the 13 languages and
the Emoji Sentiment Ranking. The correlation values, significant at the 1% level, are indicated by *.
Tweets Different Pearson Spearman rank
with emojis emojis used correlation correlation
Emoji Sent. Rank. 69,546 751 / /
English 19,819 511 0.834* 0.819*
Spanish 22,063 448 0.552* 0.573*
Polish 8,112 253 0.810* 0.783*
Russian 5,007 221 0.777* 0.756*
Hungarian 2,324 176 0.588* 0.612*
German 3,062 142 0.782* 0.783*
Swedish 2,797 139 0.702* 0.674*
Ser/Cro/Bos 2,096 123 0.708* 0.615*
Slovak 1,526 108 0.620* 0.499*
Slovenian 996 66 0.526* 0.541*
Portuguese 796 56 0.410* 0.429*
Bulgarian 607 36 0.557* 0.443*
Albanian 341 19 0.363 0.416
For each language, we form a list of emojis used in the collected tweets of the language, cut off the
emojis with fewer than 5 occurrences (the same threshold as applied to the overall Emoji Sentiment
Ranking), and compute their sentiment score. We compute the correlation coefficients between the
Emoji Sentiment Ranking and the individual languages. As can be seen in Table 5, the number of
emojis actually used in the different languages (above the threshold) drops considerably. However,
their sentiment scores and ranking remain stable. Both Pearson’s correlation and Spearman’s rank
correlation are relatively high, and significant for all the languages, except Albanian. This result is
biased towards languages with more tweets since they have a larger share in the joint Emoji Sentiment
Ranking. An alternative test might compare individual languages and the Emoji Sentiment Ranking
with the language removed. However, as a first approximation, it seems reasonable to use the Emo ji
Sentiment Ranking as a universal, language-independent resource, at least for European languages.
11

3 Conclusions
In this paper we describe the construction of an emo ji sentiment lexicon, the Emoji Sentiment
Ranking, the first such publicly available resource. We have formalized and analyzed the sentiment
properties of the emo jis in depth and highlighted some interesting conclusions.
The data that enabled these analyses, 1.6 million annotated tweets in 13 different languages, is a
valuable resource with many other useful applications. In particular, we are constructing
sentiment-classification models for different languages, and applying them to various tasks. The
Slovenian and Bulgarian language-sentiment models were already applied to monitor the mood on
Twitter during political elections in realtime [35]. The English sentiment model was used to compare
the sentiment leanings of different retweet network communities towards various environmental
topics [36]. A domain-specific English sentiment model (from another set of financial tweets) was
applied to analyze the effects of Twitter sentiment on stock prices [37]. Yet another English sentiment
model was constructed by combining a large set of general, emoticon-labeled tweets with
domain-specific financial tweets, and tested for Granger causality on the Baidu stocks [38]. The same
methodology of manual text annotations, automated model construction, and sentiment classification
was also applied to Facebook comments in Italian, where the emotional dynamics in the spreading of
conspiracy theories was studied [26].
The sentiment annotation of tweets by humans is expensive. Emoticons were already used as a
proxy for the sentiment labels of tweets. We expect that the Emoji Sentiment Ranking will turn out
to be a valuable resource for helping humans during the annotation process, or even to automatically
label the tweets with emojis for sentiment. In a lexicon-based approach to sentiment analysis, the
emoji lexicon can be used in combination with a lexicon of sentiment-bearing words. Alternatively, an
emoji with already-known sentiment can act as a seed to transfer the sentiment to the words in
proximity. Such a corpus-based approach can be used for an automated corpus construction for
feature generation [12], and then applied to train a sentiment classifier.
There are other dimensions of sentiment that are beyond the one-dimensional scale from negativity
to positivity and worth exploring. The expressiveness of the emojis allows us to assign them more
subtle emotional aspects, such as anger, happiness, or sadness, and some shallow semantics, such as
activities, locations, or ob jects of interest. An additional structuring of the emo jis can be derived from
correlations between their sentiment, e.g., various versions of hearts expressing love. However, we
consider the interplay between the emojis and the text to be one of the most promising directions for
future work. Not only the position of an emoji, but certainly its textual context is also important in
determining the role of the emoji as an amplifier and modifier of the meaning.
In the future, it will be interesting to monitor how the use of emojis is growing, and if textual
communication is increasingly being replaced by a pictorial language. Also, the sentiment and meaning
of emojis evolve over time. It might be interesting to investigate the convergence of agreement on the
meaning of controversial emojis, and to study the underpinnings of the corresponding social processes.
4 Methods
Ethics statement. The tweets were collected through the public Twitter API and are subject to the
Twitter terms and conditions. The sentiment annotations were supported by the Goldfinch platform,
provided by Sowa Labs (http://www.sowalabs.com). The human annotators were engaged for the
purpose, and were aware that their annotations will be used to construct the sentiment-classification
models, and to estimate the inter-annotator agreement and the annotator self-agreement.
4.1 Data collection
The main source of the data used in this study is a collection of tweets, in 13 European languages,
collected between April 2013 and February 2015. Most of the tweets (except English) were collected
during a joint project with Gama System (http://www.gama-system.si), using their
12

PerceptionAnalytics platform (http://www.perceptionanalytics.net). The tweets of selected
languages were collected through Twitter Search API, by specifying the geolocations of the largest
cities. For English tweets, we used Twitter Streaming API (a random sample of 1% of all public
tweets), and filtered out the English posts.
Table 6. Tweets annotated for sentiment in different languages. Languages are in
alphabetical order, Ser/Cro/Bos denotes a union of tweets in Serbian, Croatian and Bosnian.
No. of No. of
Language tweets annotators
Albanian 53,005 13
Bulgarian 67,169 18
English 103,034 9
German 109,130 5
Hungarian 68,505 1
Polish 223,574 8
Portuguese 157,393 1
Russian 107,773 1
Ser/Cro/Bos 215,657 13
Slovak 70,425 1
Slovenian 133,935 7
Spanish 275,588 5
Swedish 58,547 1
Total 1,643,735 83
We have engaged 83 native speakers (except for English) to manually annotate for sentiment over
1.6 million of the collected tweets. The annotation process was supported by the Goldfinch platform
designed specifically for sentiment annotation of short texts (such as Twitter posts, Facebook
comments, ...). The annotators were instructed to label each tweet as either negative,neutral, or
positive, by estimating the emotional attitude of the user who posted the tweet. They could also skip
the inappropriate or irrelevant tweets. The breakdown of the annotated tweets by language is in
Table 6.
Another source of data comes from Emo jitracker (http://emojitracker.com/). Emojitracker
monitors and counts the number of emojis used on Twitter in realtime. It has been active since July
2013, and so far it has detected over 10 billion emoji occurrences. We downloaded the current count of
emoji occurrences as of June 2015. This data is used to estimate how representative is our sample of
emojis in the annotated tweets.
The data from both sources is available in a public language-resource repository clarin.si at
http://hdl.handle.net/11356/1048. There are two data tables, in an open csv format, one for the
Emoji Sentiment Ranking, and the other from Emojitracker. The tables list all the emojis found, their
occurrences, and, in the case of the Emoji Sentiment Ranking, also their numbers in the negative,
neutral, and positive tweets. From this data, the Emoji Sentiment Ranking web page at
http://kt.ijs.si/data/Emoji_sentiment_ranking/ is automatically generated.
4.2 Emoji Unicode symbols
The exact definition of what constitutes an emoji symbol is still emerging. In particular, there is some
discrepancy between our set of emojis and the emojis tracked by Emojitracker. Also, during the
writing of this paper, in August 2015, the Unicode consortium published a new set of emojis, the
Unicode Emo ji Charts (http://www.unicode.org/emoji/).
The set of emojis in our Emoji Sentiment Ranking follows the Unicode standard version 8 [15] and
consists of all the single-character symbols from the Unicode category ‘Symbol, Other’ (abbreviated
[So]) that appear in our tweets. Emojitracker, on the other hand, also tracks some double-character
symbols (10 Country Flags, and 11 Combining Enclosing Keycaps), but does not track all the [So]
13

symbols that appear in our data. In particular, 49 Dingbats, 46 Miscellaneous Symbols, 38 Box
Drawings, 28 Geometric Shapes, 21 Enclosed Alphanumerics, 20 Enclosed Alphanumeric Supplements,
and 13 Arrows are not tracked. The Unicode Emo ji Charts have introduced even more new emoji
symbols, in particular an exhaustive list of 257 double-character Country Flags. A comparison of the
overlaps and differences in the emoji symbol specifications between the three sources is in Tables 7
and 8.
Table 7. Types and numbers of emo ji symbols. [So] is an abbreviation for the Unicode
category ‘Symbol, Other’.
No. of single [So]
all character non-[So]
emoji double flags
symbols character keycaps
Emoji Sentiment Ranking 969 969 969
0
0 0
0
Emojitracker 845 824 812
12
21 10
11
Unicode Emoji Charts 1281 1012 995
17
269 257
12
Table 8. Overlaps and differences for emojis from the three data sources. A table entry is
the number of emojis in (∈), or missing ( /∈) from a data source. N(S ingle, Double) denotes the total
number Nof emoji symbols, partitioned into the Single- and Double-character symbols, respectively.
Emoji Sentiment Ranking
∈/∈Total
Emojitracker ∈721 (721, 0) 124 (103, 21) 845 (824, 21)
/∈248 (248, 0) / /
Unicode ∈734 (734, 0) 547 (278, 269) 1281 (1012, 269)
Emoji Charts /∈235 (235, 0) / /
Total 969 (969, 0) / /
The emoji symbols that are not common to all the three data sources are relatively infrequent. The
highest-ranking emoji in Emojitracker, which is absent from our data, has the rank 157 (double
exclamation mark). The highest-ranking emoji in the Emoji Sentiment Ranking, not tracked by
Emojitracker, has the rank 13 (white heart suit). Additionally, we noticed that we missed three
characters from the [So] category: ‘degree sign’, ‘numero sign’, and ‘trade mark sign’. However, only
the ‘trade mark sign’ (with 257 occurrences in our data) is also considered by the Emojitracker and
the Unicode Emoji Charts. Despite these minor differences in the emoji sets, all our results remain
valid. However, in the next version of the Emo ji Sentiment Ranking we plan to extend our set to
double-character symbols, and consider all the emojis from the Unicode Emoji Charts as an
authoritative source.
14

4.3 Sentiment formalization
The sentiment of an individual tweet can be negative,neutral, or positive. Formally, we represent it by
a discrete, 3-valued variable, c, which denotes the sentiment class:
c∈ {−1,0,+1}
This variable models well our assumptions about the ordering of the sentiment values and the
distances between them.
An object of Twitter posts to which we attribute sentiment (an emoji in our case, but it can also
be a stock [37], a political party [35], a discussion topic [26, 36], etc.) occurs in several tweets. A
discrete distribution:
N(c),X
c
N(c) = N , c ∈ {−1,0,+1},
captures the sentiment distribution for the set of relevant tweets. Ndenotes the number of all the
occurrences of the object in the tweets, and N(c) are the occurrences in tweets with the sentiment
label c. From the above we form a discrete probability distribution:
(p−, p0, p+),X
c
pc= 1 .
For convenience, we use the following abbreviations:
p−=p(−1) , p0=p(0) , p+=p(+1) ,
where p−,p0, and p+denote the negativity,neutrality, and positivity of the object (an emoji in
our case), respectively. In SentiWordNet [16], the term objectivity is used instead of the neutrality
p0. The sub jectivity can then be defined as p−+p+[39].
Typically, probabilities are estimated from relative frequencies, pc=N(c)/N. For large samples,
such estimates are good approximations. Often, however, and in particular in our case, we are dealing
with small samples, e.g., N= 5. In such situations, it is better to use the Laplace estimate (also
known as the rule of succession) to estimate the probability [40]:
pc=N(c) + 1
N+k,(for large N:pc≈N(c)
N).
The constant kin the denominator is the cardinality of the class, in our case k=|c|= 3. The Laplace
estimate assumes a prior uniform distribution, which makes sense when the sample size is small.
Once we have a discrete probability distribution, with properly estimated probabilities, we can
compute its mean:
¯x=X
c
pc·c .
We define the sentiment score, ¯s, as the mean of the discrete probability distribution:
¯s=−1·p−+ 0 ·p0+ 1 ·p+=p+−p−.
The sentiment score has the range: −1<¯s < +1.
The standard deviation of a discrete probability distribution is:
sd =sX
c
pc·(c−¯x)2,
and the standard error of the mean is:
sem =sd
√N.
15

4.4 Sentiment bar
The sentiment bar is a useful, novel visualization of the sentiment attributed to an emoji (see
http://kt.ijs.si/data/Emoji_sentiment_ranking/ for examples). In a single image, it captures
all the sentiment properties, computed from the sentiment distribution of the emoji occurrences:
p−, p0, p+,¯s, and ¯s±1.96sem (the 95% confidence interval). Three examples that illustrate how the
sentiment properties are mapped into the graphical features are shown in Fig 9. The top sentiment
bar corresponds to the ‘thumbs down sign’ emoji, and indicates negative sentiment, with high
confidence. The middle bar represents the estimated sentiment of the ‘flushed face’ emoji. The
sentiment is neutral, close to zero, where both negative and positive sentiment are balanced. The
bottom bar corresponds to the ‘chocolate bar’ emoji. Its sentiment score is positive, but its standard
error bar extends into the neutral zone, so we can conclude with high confidence only that its
sentiment is not negative.
CHOCOLATE BAR
FLUSHED FACE
THUMBS DOWN SIGN
1.96 SEM
-1 0 +1
pp0p+
s
Figure 9. Sentiment bars of the ‘thumbs down sign’, ‘flushed face’, and ‘chocolate bar’
emojis. The colored bar extends from −1 to +1, the range of the sentiment score ¯s. The grey bar is
centered at ¯sand extended for ±1.96sem, but never beyond the range of ¯s. Colored parts are
proportional to negativity (p−, red), neutrality (p0, yellow), and positivity (p+, green).
4.5 Welch’s t-test
Welch’s t-test [31] is used to test the hypothesis that two populations have equal means. It is an
adaptation of Student’s t-test, but is more reliable when the two samples have unequal variances and
sample sizes. Welch’s t-test is also robust for skewed distributions and even more for large sample
sizes [32].
Welch’s t-test defines the t statistic as follows:
t=¯x1−¯x2
qsd2
1
N1+sd2
2
N2
.
The degrees of freedom, ν, are estimated as follows:
ν≈



(sd2
1
N1+sd2
2
N2)2
sd4
1
N2
1(N1−1) +sd4
2
N2
2(N2−1) 


,
16

where ⌊⌋ denotes the approximate degrees of freedom, rounded down to the nearest integer. Once the
t value and the degrees of freedom are determined, a p-value can be found from a table of values for
Student’s t-distribution. For large degrees of freedom, ν > 100, the t-distribution is very close to the
normal distribution. If the p-value is below the threshold of statistical significance, then the null
hypothesis is rejected.
4.6 Pearson and Spearman correlations
We need to correlate two properties of the Emoji Sentiment Ranking with other data. In the first case
we correlate the emojis ranked by occurrence to the Emojitracker list—the property of the list
elements is the number of occurrences. In the second case we correlate the emojis ranked by sentiment
to subsets of emojis from the 13 different languages—the property of the list elements is the sentiment
score.
For any two lists xand y, of length n, we first compute the Pearson correlation coefficient [29]:
r(x, y) = Pn
i=1 (xi−¯x)(yi−¯y)
pPn
i=1 (xi−¯x)2Pn
i=1 (yi−¯y)2,
where ¯xand ¯yare the list’s mean values, respectively. The Spearman’s rank correlation coefficient [30]
is computed in the same way, the property values of the xand yelements are just replaced with their
ranks. In both cases we report the correlation coefficients at the 1% significance level.
4.7 Agreement measures
In general, an agreement can be estimated between any two methods for generating data. In our case
we want to estimate the agreement between humans when annotating the same tweets for sentiment.
A comparison of agreements between different datasets gives some clue about how difficult the task is.
There are different measures of agreement, and to get a robust estimate of the differences, we apply
three well-known measures.
Krippendorff’s Alpha-reliability [33] is a generalization of several specialized agreement measures.
It works for any number of annotators, is applicable to different variable types and metrics (e.g.,
nominal, ordered, interval, etc.), and can handle small sample sizes. Alpha is defined as follows:
Alpha = 1 −Do
De
,
where Dois the observed disagreement between the annotators, and Deis the disagreement expected
by chance. When the annotators agree perfectly, Alpha = 1, and when the level of agreement equals
the agreement by chance, Alpha = 0. The two disagreement measures are defined as follows:
Do=1
NX
c,c′
N(c, c′)·δ2(c, c′),
De=1
N(N−1) X
c,c′
N(c)·N(c′)·δ2(c, c′).
The arguments, N, N (c, c′), N (c), and N(c′), refer to the frequencies in a coincidence matrix, defined
below. δ(c, c′) is a difference function between the values of cand c′, and depends on the metric
properties of the variable. In our case, for the discrete sentiment variables cand c′, the difference
function δis defined as:
δ(c, c′) = |c−c′|c, c′∈ {−1,0,+1}.
In [33], this is called the interval difference function. Note that the function attributes a disagreement
of 1 between the negative (or positive ) and the neutral sentiment, and a disagreement of 2 between the
negative and positive sentiments.
17

Acoincidence matrix tabulates all the pairable values of cfrom two different annotators into a
k-by-ksquare matrix, where k=|c|. Unlike a contingency matrix (used in association and correlation
statistics) which tabulates pairs of values, a coincidence matrix tabulates all the pairable values. A
coincidence matrix omits references to annotators. It is symmetrical around the diagonal, which
contains all the perfect matches. A coincidence matrix has the following general form:
c′P
. . .
c . N(c, c′). N(c)
. . .
PN(c′)N
Here cand c′range over all possible values of the variable. In a coincidence matrix, each labeled unit
is entered twice, once as a (c, c′) pair, and once as a (c′, c) pair. N(c, c′) is the number of units labeled
by the values cand c′by different annotators, N(c) and N(c′) are the totals for each value, and Nis
the grand total. Note that Nis two times the number of units labeled by the different annotators.
In the case of sentiment annotations, we have a 3-by-3 coincidence matrix. Two example matrices
are shown in Tables 9 and 10. Note that both coincidence matrices in Tables 9 and 10 are symmetric
around the diagonal, and that the totals Nare two times larger than in Table 3 because each
annotated tweet is counted twice.
Table 9. Coincidence matrix for tweets with emojis.
Sentiment Negative Neutral Positive Total
Negative 1,070 354 196 1,620
Neutral 354 902 725 1,981
Positive 196 725 2,572 3,493
Total 1,620 1,981 3,493 7,094
Table 10. Coincidence matrix for tweets without emojis.
Sentiment Negative Neutral Positive Total
Negative 15,356 7,777 3,004 26,137
Neutral 7,777 23,670 10,921 42,368
Positive 3,004 10,921 21,624 35,549
Total 26,137 42,368 35,549 104,054
In machine learning, a classification model is automatically constructed from the training data and
evaluated on a disjoint test data. A common, and the simplest, measure of the performance of the
model is Accuracy , which measures the agreement between the model and the test data. Here, we use
the same measure to estimate the agreement between the pairs of annotators. Accuracy is defined in
terms of the observed disagreement Do:
Accuracy = 1 −Do=1
NX
c
N(c, c).
Accuracy is simply the fraction of the diagonal elements of the coincidence matrix. Note that it does
not account for the (dis)agreement by chance, nor for the ordering between the sentiment values.
Another, more sophisticated measure of performance, specifically designed for 3-class sentiment
classifiers [12], is F1(−,+):
F1(−,+) = F1(−) + F1(+)
2.
F1(−,+) implicitly takes into account the ordering of the sentiment values by considering only the
negative (−) and positive (+) labels, and ignoring the middle, neutral label. In general, F1(c) (known
18

as the F-score) is a harmonic mean of precision and recall for class c. In the case of a coincidence
matrix, which is symmetric, the ‘precision’ and ‘recall’ are equal, and thus F1(c) degenerates into:
F1(c) = N(c, c)
N(c).
In terms of the annotator agreement, F1(c) is the fraction of equally labeled tweets out of all the
tweets with label c.
Acknowledgments
This work was supported in part by the EC projects SIMPOL (no. 610704), MULTIPLEX (no.
317532) and DOLFINS (no. 640772), and by the Slovenian ARRS programme Knowledge Technologies
(no. P2-103).
We acknowledge Gama System (http://www.gama- system.si) who collected most of the tweets
(except English), and Sowa Labs (http://www.sowalabs.com) for providing the Goldfinch platform
for the sentiment annotation of the tweets. We thank Saˇso Rutar for generating the Emoji Sentiment
Ranking web page, Andrej Blejec for statistical insights, and Vinko Zlati´c for suggesting an emoji
distribution model.
References
1. Dimson T. Emo jineering part 1: Machine learning for emoji trends [blog]; 2015.
http://instagram- engineering.tumblr.com/post/117889701472/emojineering-part-1-machine-learning-for-
2. SwiftKey PT. Most-used emoji revealed: Americans love skulls, Brazilians love cats, the French
love hearts [blog]; 2015.
http://swiftkey.com/en/blog/americans- love-skulls-brazilians-love-cats-swiftkey-emoji-meanings-rep
3. Liu B. Sentiment Analysis and Opinion Mining. Synthesis Lectures on Human Language
Technologies. 2012;5(1):1–167. Available from:
http://dx.doi.org/10.2200/S00416ED1V01Y201204HLT016.
4. Liu B. Sentiment Analysis: Mining Opinions, Sentiments, and Emotions. Cambridge University
Press; 2015.
5. Boia M, Faltings B, Musat CC, Pu P. A :) is worth a thousand words: How people attach
sentiment to emoticons and words in tweets. In: Intl. Conf. on Social Computing (SocialCom).
IEEE; 2013. p. 345–350.
6. Hogenboom A, Bal D, Frasincar F, Bal M, de Jong F, Kaymak U. Exploiting emoticons in
polarity classification of text. Journal of Web Engeneering. 2015;14(1-2):22–40.
7. Hogenboom A, Bal D, Frasincar F, Bal M, de Jong F, Kaymak U. Exploiting emoticons in
sentiment analysis. In: Proc. 28th Annual ACM Symposium on Applied Computing. ACM;
2013. p. 703–710.
8. Davidov D, Tsur O, Rappoport A. Enhanced sentiment learning using Twitter hashtags and
smileys. In: Proc. 23rd Intl. Conf. on Computational Linguistics: Posters. ACL; 2010. p.
241–249.
9. Liu KL, Li WJ, Guo M. Emoticon smoothed language models for Twitter sentiment analysis.
In: Proc. 26th AAAI Conf. on Artificial Intelligence; 2012. p. 1678–1684.
10. Read J. Using emoticons to reduce dependency in machine learning techniques for sentiment
classification. In: Proc. ACL Student Research Workshop. ACL; 2005. p. 43–48.
19

11. Zhao J, Dong L, Wu J, Xu K. Moodlens: An emoticon-based sentiment analysis system for
Chinese tweets. In: Proc. 18th ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data
Mining. ACM; 2012. p. 1528–1531.
12. Kiritchenko S, Zhu X, Mohammad SM. Sentiment analysis of short informal texts. Journal of
Artificial Intelligence Research. 2014;p. 723–762.
13. The Unicode Consortium, Allen JD, et al. The Unicode Standard, Version 6.0. Mountain View,
CA; 2011. Available from: http://www.unicode.org/versions/Unicode6.0.0/.
14. The Unicode Consortium, Allen JD, et al. The Unicode Standard, Version 7.0. Mountain View,
CA; 2014. Available from: http://www.unicode.org/versions/Unicode7.0.0/.
15. The Unicode Consortium, Allen JD, et al. The Unicode Standard, Version 8.0. Mountain View,
CA; 2015. Available from: http://www.unicode.org/versions/Unicode8.0.0/.
16. Baccianella S, Esuli A, Sebastiani F. SentiWordNet 3.0: An enhanced lexical resource for
sentiment analysis and opinion mining. In: LREC. vol. 10; 2010. p. 2200–2204.
17. Miller GA. WordNet: A lexical database for English. Communications of the ACM.
1995;38(11):39–41.
18. Go A, Bhayani R, Huang L. Twitter sentiment classification using distant supervision. CS224N
Project Report, Stanford. 2009;1:12.
19. Pak A, Paroubek P. Twitter as a corpus for sentiment analysis and opinion mining. In: LREC.
vol. 10; 2010. p. 1320–1326.
20. Thelwall M, Buckley K, Paltoglou G, Cai D, Kappas A. Sentiment strength detection in short
informal text. Journal of the American Society for Information Science and Technology.
2010;61(12):2544–2558.
21. Amalanathan A, Anouncia SM. Social network user’s content personalization based on
emoticons. Indian Journal of Science and Technology. 2015;8(23).
22. Ganesan KA, Sundaresan N, Deo H. Mining tag clouds and emoticons behind community
feedback. In: Proc. 17th Intl. Conf. on World Wide Web. ACM; 2008. p. 1181–1182.
23. Kramer AD. The spread of emotion via Facebook. In: Proc. SIGCHI Conf. on Human Factors
in Computing Systems. ACM; 2012. p. 767–770.
24. Kramer AD, Guillory JE, Hancock JT. Experimental evidence of massive-scale emotional
contagion through social networks. Proc. National Academy of Sciences.
2014;111(24):8788–8790.
25. Coviello L, Sohn Y, Kramer AD, Marlow C, Franceschetti M, Christakis NA, et al. Detecting
emotional contagion in massive social networks. PLoS ONE. 2014;9(3):e90315.
26. Zollo F, Novak Kralj P, Del Vicario M, Bessi A, Mozetiˇc I, Scala A, et al. Emotional dynamics
in the age of misinformation. PLoS ONE. 2015;10(9):e138740. Available from:
http://dx.doi.org/10.1371/journal.pone.0138740.
27. Gruzd A, Doiron S, Mai P. Is happiness contagious online? A case of Twitter and the 2010
Winter Olympics. In: Proc. 44th Hawaii Intl. Conf. on System Sciences (HICSS). IEEE; 2011.
p. 1–9.
28. Hatfield E, Cacioppo JT, Rapson RL. Emotional Contagion. Cambridge University Press; 1994.
20

29. Pearson K. Note on regression and inheritance in the case of two parents. Proceedings of the
Royal Society of London. 1895;58:240–242.
30. Spearman C. The proof and measurement of association between two things. The American
Journal of Psychology. 1904;15:72–101.
31. Welch BL. The generalization of ”Student’s” problem when several different population
variances are involved. Biometrika. 1947;34(1-–2):28––35.
32. Fagerland MW. t-tests, non-parametric tests, and large studies—a paradox of statistical
practice? BMC Medical Research Methodology. 2012;12(78).
33. Krippendorff K. Content Analysis, An Introduction to Its Methodology. 3rd ed. Thousand
Oaks, CA: Sage Publications; 2012.
34. Newman MEJ. Power laws, Pareto distributions and Zipf’s law. Contemporary Physics.
2005;46(5):323–351. Available from: http://arxiv.org/abs/cond-mat/0412004.
35. Smailovi´c J, Kranjc J, Grˇcar M, ˇ
Znidarˇsiˇc M, Mozetiˇc I. Monitoring the Twitter sentiment
during the Bulgarian elections. In: Proc. IEEE Intl. Conf. on Data Science and Advanced
Analytics (DSAA). IEEE; 2015.
36. Sluban B, Smailovi´c J, Battiston S, Mozetiˇc I. Sentiment leaning of influential communities in
social networks. Computational Social Networks. 2015;2(9). Available from:
http://dx.doi.org/10.1186/s40649- 015-0016-5.
37. Ranco G, Aleksovski A, Caldarelli G, Grˇcar M, Mozetiˇc I. The effects of Twitter sentiment on
stock price returns. PLoS ONE. 2015;10(9):e138441. Available from:
http://dx.doi.org/10.1371/journal.pone.0138441.
38. Smailovi´c J, Grˇcar M, Lavraˇc N, ˇ
Znidarˇsiˇc M. Stream-based active learning for sentiment
analysis in the financial domain. Information Sciences. 2014;285:181–203. Available from:
http://dx.doi.org/10.1016/j.ins.2014.04.034.
39. Zhang W, Skiena S. Trading strategies to exploit blog and news sentiment. In: Proc. 4th Intl.
AAAI Conf. on Weblogs and Social Media; 2010. p. 375-378.
40. Good IJ. The Estimation of Probabilities: An Essay on Modern Bayesian Methods. Cambridge,
MA: MIT Press; 1965.
21