Cher is the Queen of Emoji even if she isn’t [Data analysis]

It is universally recognized by experts that Cher is the Queen of Emoji. (Hail, Cher.)

I’m pretty sure this is what Cher wears while she tweets emoji after emoji after emoji

Style👡 and a sense📡 of the 👑💃🏽queen💃🏽👑

Let’s get started with a Cher tweet that I’ve chosen almost-at-random:

  • Length — These days, Cher’s tweets are consistently among the longest; Cher’s tweets are around 137 characters, while DJ Khaled’s are 79, Chelsea Manning’s are typically 55 and Kyle MacLachlan’s are 56. Across 55 celebrities, a median length of 73 characters is typical. Among all 55 celebrities, Cher consistently packs in the most characters.
  • Spacing — Cher routinely cuts out unnecessary spaces like the ones after punctuation
  • Homophonous abbreviation — this is just my fancy way of saying she uses “2” for “to”
  • Dropped apostrophes — so far in 2017 Cher has 18 he’s and 75 hes, same basic ratios for she’s/shes and i’m/im
  • 🔥 — this is one of the most popular emojis for celebrities to use, though here, Cher seems to be using it as a substitute for the adjective flaming. Consider uses by Nicki Minaj, Snoop Dogg, DJ Khaled, Chad Johnson, Chelsea Peretti, Chelsea Manning, Kyle MacLachlan, and William Shatner. Those are all fairly expressive “paralinguistic” uses. But Cher is more likely than most others to treat 🔥 as a word. You can see a bit of something similar here in Lena Dunham and Dwayne Johnson’s 🔥=on fire or Alyssa Milano’s semi-ornamental 🔥=terminate from employment. Though I haven’t seen others do what Cher does here where she adds a past tense marker to talk about Republicans fiddling while Rome 🔥🔥’d. Wordiness!
  • 🐝 — Cher commonly uses this as a substitute for the verb be. This is the same basic thing thing as with 2 above, homophonous abbreviation. Note that there isn’t any saved effort. Cher usually works on an iPhone and you have to press one key to get to the numbers or emoji and then a second key to get them to appear. Two key strokes, the same number of letters as to and be. More likely, you do this to get something more visually interesting. Emoji tend to be a way of playing with language and here Cher is calling back an old form of visual linguistic play — the rebus puzzle. (Aside: maybe that’s what William Shatner is doing here?!)
  • 🤡 — Hey, the best clowns are down in the sewers. Among the 55 celebrities I’ve looked at, the biggest users are Cher (28 times), Don Cheadle (21 times) and Jeffrey Wright (9 times). They mostly use it within tweets about Donald Trump, as Cher is doing in the tweet shown above. There do seem to be a smattering of affectionate uses of the clown by other folks, but mostly clown=bad instead of a happy Red Skelton kinda thing.
In 2017, Cher’s tweets are among the longest of celebrities. But this, what could it mean!? (Answer: she was getting her star on the Hollywood Walk of Fame) This is a time before anyone could really use emoji, btw (see the next section for more history/timelines).
Cher loves exclamation points, the big red double !! is actually one of the main ones she adopts later on. Cher also loves ALL CAPS, which you see elsewhere in what she calls her rant. This particular thread was kicked off by Cher hearing a conservative jerk talk about gay people; fwiw, the rant was 9 tweets over 14 min…although there were a number of follow ups about an hour later
One of my favorite things in life is people who end text messages or tweets with “Bye”.
Ah, remember Vines? They got disabled in October 2016.
Cher LOVES these emoji and uses them more than pretty much anyone else. Notice her use of line breaks, too. And, of course, a deep dislike of Donald Trump. That is a hallmark of her tweets in recent years, as you’re about to see.
This is quintessential recent Cher — fairly long, lots of ALL CAPS, some emoji and a predilection to ignore spaces after punctuation marks
Cher’s all-time most-favorited tweet. Have I mentioned that she doesn’t like Donald Trump?
  • Christine Teigen also uses a lot of verbs (am, wondering), but her other relative favorite is adverbs (basically, literally)
  • Donald Trump uses a lot of nouns (business, media), prepositions (including, after), adjectives (fake, great), and articles (the, an)
  • Ellen Degeneres uses a lot of pronouns (something, her). So does Penn Jillette (we, nothing), but he also likes conjunctions (but, and).
  • Joyce Carol Oates uses a lot of adjectives (public, liberal) and negations
  • Jenny Slate loves adverbs (just, also), determiners (this, many) and interjections (boo, bye, oh)

A history of emoji through Cher 📜📱

The very first emoji appeared on Japanese cellphones in 1999, created by Shigetaku Kurita, the first 176 emoji would only show up if you and the people you were texting both had NTT Docomo devices. Today, we have Unicode so that emoji appear pretty much anywhere, but it took about a dozen years from the creation of emoji in Japan til worldwide use.

The very first emoji (psst: notice how few emotions there are)

The queen, DEPOSED, 💔

I’ve tried to put this off, but it is time for a reckoning.

  • Kyle MacLachlan: 👍, 🎂, 😀, 🍷, 🎉
  • Chelsea Manning: 😎, 🌈, 💕, 😇, 😌
  • DJ Khaled: 🌺 , 🙏, 🏽, 🔑, 🦁
Fwiw, people generally do put faces before other symbols
Although my dissertation was partly on emoticons, I still feel 🦄 about this tweet
You should click the link to see some good forest poppn
You thought the 😤 was about steam coming out of someone’s nose because they were angry. Adaptation!
Snoop Dogg comes out on top whether we choose sums, averages, etc. He posts to Twitter via Instagram a lot — check out this post on emoji-on-Instagram; Chelsea Manning didn’t have an account for most of this time period, so she’s at a distinct disadvantage in this particular chart. See the very last graphic in this post for 55 celebrities’ usage in 2017 only (she’s in 7th place).
Knives + drugs + eggplants: not my favorite

📈 or 📉: Are we past peak (celebrity) emoji usage?

If you look at rates of emoji use across the years, you do see what looks like a peak for a lot of celebrities last year (2016). And there are some plummets since 2016 — DJ Khaled and Kyle MacLachlan aren’t emojifying their posts as much. But Snoop Dogg and Cher are holding steady and the majority are increasing: Nicki Minaj, Katy Perry, Diplo, William Shatner, Chad Johnson (ochocinco), and Chelsea Manning (xychelsea).

Do you believe?💋🎶☝

In normal speech, we’ve got all kinds of cues to help interpret messages. But when you switch to written forms, you lose intonation, volume, furrowed brows, dodged eye contact and the like. Historically, lots of writing hasn’t needed these things (quick, page three of your legal contract should have more eye rolling). But recent years have seen an explosion of writing in the form of SMS text messaging and social media. Emoji have emerged since ~2012 as a viable augmentation, a kind of adornment that helps soften messages that might sound monotone/harsh as well as a way for people to play with language.

End notes for the curious 🤔

  1. Cher has tweeted 489,369 unigrams since 2010 — that counts stuff like punctuation; if you want “words” as in “stuff in dictionaries with parts of speech”, then it’s more like 78,108 words since 2010. That ignores URLs, hashtags, @’ing people, super-slang, numbers, punctuation, emoticons, emoji, etc. Cher is not quite in the top quartile when it comes to the ratio of these numbers. Diplo and Nicki Minaj have a lot more normal words, Kyle MacLachlan and Chelsea Manning have a lot fewer.
  2. For 2017, Cher used 5,355 base emoji, she paired them with gender 7 times, she paired them with skin color 533 times, and she had 224 flags (technically, flags tend to be made up of two emoji characters each…so 224 is actually the total divided by two). It could make sense to treat “very white female skier” as one emoji, but I count it as three. That also helps me understand how people are deploying things like gender and race better.
  3. Re per million words: Normally people use “wpm” not “pmw”. And while I like the idea of following the general trend, I think “per million words” is closer to what the little voice in our heads should say as we read along here than “words per million”.
  4. How did I pick the 55 celebrities? Well, I actually had lots more but limited myself to people over 750 tweets in 2017 (152 other celebrities didn’t make the cut). I chose 750 tweets to make sure that Kyle MacLachlan was included since he is emoji legendary. The overall list was assembled from looking across a lot of articles where celebrities were mentioned in terms of emoji and/or Twitter uses. That got me, for example, a bunch of athletes I frankly hadn’t heard of (sorry! please get a Grammy, Oscar, and Emmy like Cher has). I then looked at who all these people @’ed the most and if they were (vaguely) famous, I put them in, too. So it’s pretty eclectic. For pre-2017, I only looked at the top ten emoji users from 2017, see above. See the bottom of the post for who all the 55 celebrities are.
  5. Note that skin tones are also among the most popular emoji in this data — there are five of these defined by Unicode, the ones most frequent in the data here are 🏾, 🏻, and 🏽. Since people tend to use skin tones to refer to themselves, this mostly reflects the makeup of the celebrities reported here. Of the 55 celebrities, somewhere between 40–50% of them are People of Color (I’m not going to be precise because I don’t actually know how everyone identifies). I was just going for big Twitter users who were relatively famous, so I didn’t attempt to, say, get equal numbers across the Fitzpatrick scale nor did I (to flip it around) start with skin tone modifiers and try to get equal numbers of people using them. There’s some really really interesting work left to be done on race, identification, and emoji. Especially because people don’t always use the same exact tone for themselves and it’s probably interesting (likely depressing) to see how people use skin tones when they are creating emoji that are about people other than themselves.
  6. Linguists sometimes use something called “TTR” (Type:Token Ratio) to measure vocabulary richness. To get a sense of this, continue two people, each of whom has used 100 emoji total. The idea for TTR is to award someone who uses 100 different emoji one time each a higher score than someone who uses one emoji but uses it 100 times (100/100 > 1/100). Because diversity of vocabulary increases with text length, you need to make sure you control for that by having consistent sampling. Grabbing 500 random tweets from everyone in 2017, we see that Snoop Dogg > Cher > Nicki Minaj. But even higher scores get racked up by Tyler Oakley, Diplo, Bayley, and Kyle MacLachlan.

Linguistics and language, data science and artificial intelligence, UX and design, travel, San Francisco. Want to talk about emoticons, emoji or AI?