cial Media

Made By William Zhang.

Emoji, developed from the 1990s, has become a complete and well-supported system today. We use emoji, sometimes to support our texts and express emotions, and sometimes, we spam emoji just to show the strong feeling within us. Emoji are so widely used, but do we understand emoji enough?

Emoji are often misunderstood. This misunderstanding relates to the cultural or contextual interpretation of the emoji – when the author picks an emoji, they think about it in a certain way, but the same character may not trigger the same thoughts in the mind of the receiver.

🙂 → 😑

[A smiley face could be sent to convey
a despising, mocking, even obnoxious attitude.]

Researchers found that 40 percent of emoji tested had a “sentiment misconstrual score” of 2 or more, so when two people look at the same emoji, the sentiment disagreement between them would be much greater than that. Now the question comes, are you sure the emoji sent by you means exactly the same in the recipient's mind? Or are you certain how you perceive one emoji is what the majority does?

So, the data I found is a raw data set of 18 million unprocessed English tweets with at least one emojis, as well as a file about the how these emojis are used in Twitter: how many are used in positive, neutral and negative tweets, and base on that, I could calculate the sentiment score for each emoji – if it is used more in negative tweets, then it is more negative than positive.

Let's see what the data tells us ——

Neutrality - Neutral Occurrences ÷ Total Occurrences

Sentiment Score - (Positive Occurrences - Negative Occurrences) ÷ Total Occurrences

• The higher the emoji, the more neutral.

• Emojis on the right are more positive.

• The size shows the relative occurrences.

More emojis are clustering on the right hand side, which shows there are more positive emojis (of course!), or to say, there are more positive tweets sent out each day.

[Please Double-click on the graph to zoom out.]

The emojis with more occurrences are all gathering in the lower middle part, it seems the most popular emojis are all used for expressing emotions.

[Explore yourself by selecting an area on the graph.
Hover on the emojis to show detailed data.]

Most likely, you have encountered some emojis beyond your expectation - some seemingly neutral emojis are used mostly in negative tweets??

The above are general cases, what about emojis in individual texts?


How many times do you spam emojis in text every day? Surely a lot.

Among the 6,000,000 texts processed,
1,011,166 texts contains several spamming of one emoji!

And how many times is one emoji spammed in one text?

Let's See ——

Here are some randomly chosen emojis. The dots' radius represents the number of times the emoji is spammed in one text.

[Please wait until static...]

Show some other commonly spammed emojis.

List the top most spammed emojis.

It is surely "Face With Tears of Joy😂", with the record of 205795 spams in the 6,000,000.

Following:2.😭72205; 3.🔥65277; 4.😍36792; 5.❤36695; 6.👏26464; 7.🤣20125; 8.👉13115; 9.💕13086; 10.✅12335;

The Best Spammer in my dataset spams an astounding 126 number of EMOJIS in one text.

126 times of 🙄!! That guy must be really annoyed...

With words, it is hard to directly express certain emotions, but with the strong visual cue of emojis, we could make sure our strong emotions be delivered 😜😜😜😜😜😜😜.


Where in your text do you often place your emojis? Probably in most of the times, it is more subconscious than intentional.

Could there be some pattern to find?

[Left: the beginning of tweets,
Right: the end of tweets.]

On average, an emoji is placed at 2/3 of the length of a tweet.

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.


In one text, what do you expect to combine the usage of different emojis? I randomly selected 2000 tweets out of the 6,000,000 and filtered it with unique combinations with at least 3 different emojis.

Hover to figure out!

Surely, some stories could be found beneath the emojis, or are they unexpected? ——

😩🤣❤️: "We literally drive eachother crazy..but only cause were crazy about eachother😩🤣❤️".

🌆👉🚖📞: "RT: 🌆 👉🚖 For Taxi

💪🏻🇺🇸🦅: "Army Vet Earl Granville carried a woman across the finish line of the Boston Marathon with one leg 💪🏻🇺🇸🦅 …".

Text Sentiment vs. Emoji Sentiment

Critics of emoji have expressed that:
As we outsource our feelings to tiny pictures, it narrows our ability to express ourselves honestly.
When we use emojis, we’re not revealing our emotions—we’re often stifling them.

So, Are we feeling less of the feels, while becoming more emotional on the social media?

"In fact, we don’t even have emotions—we have feels. But maybe not all the feels."

I used the Natural Language Token and VADER (Valence Aware Dictionary and sEntiment Reasoner) library of Python to give the text sentiment scores separately from emojis and see how emoji is contributing to the emotion of the tweets.

I bet the graph is a bit unexpected? From the positive graph, it seems in most times, our emotion in our text conform with the emojis' emotion.
However, from the negative graph, most emojis' sentiment score are positive?
That opens a new topic beyond my reach, but we may guess that when we are expressing negative emotions, we tend to use positive emojis to cover it up?

Text Sentiment

Emoji Sentiment

"Oof, sounds rough."
"Wait, what? That’s a smile…"
"No, it’s a grimace."

Emojis—🎉😎🚀🍕—began as a light-hearted🤍 way to breathe🌬️ life into our interactions🙆‍, and have since evolved into a creative art form🎨.

Now that our lives live online🌐, we are presented with daily conundrums at a dizzying pace💫.

When a friend’s dog🐕 dies, do we choose a sad reaction😥 on Facebook or repeat a platitude🙏🏻 that 10 others have already posted?

When your best friend says they’re feeling blue☄️, do you spend your limited bandwidth⌛ figuring out what to say🙊, or do you choose an appropriate gif🌆 to do it for you?

When an acquaintance bares their soul about a rocky⛰️ divorce, does it warrant a heart💖 or angry face👿?

Or do you just skim👀 past the post entirely, and blame the algorithm🧮 if they question your inaction🙈?



The End.❤

Made By William Zhang.

Data Sources:

EmojifyData-EN: English tweets, with emojis: https://www.kaggle.com/rexhaif/emojifydata-en

Emoji Sentiment Ranking 1.0: https://www.clarin.si/repository/xmlui/handle/11356/1048