If I send you a smiley or frowny emoji—like or —it's clear that it signifies a positive or negative emotion. But what about , , or ? To most people, they immediately convey some kind of emotional content. But to a computer, they're just strings of numbers and letters that get translated into images by Web browsers. That's a problem for researchers who want to quickly classify the emotional content of online communication. The solution? A Rosetta Stone that translates these emojis into general emotions. Using 83 human subjects fluent in a total of 13 European languages, researchers have built their first prototype, the Emoji Sentiment Ranking. The database classifies 751 of the most frequently used emojis as negative, neutral, or positive. By using the new database to analyze tweets in all 13 languages, the researchers found that what we're blurting out on the Internet is mostly upbeat. Less than 20% of tweeted emojis are negative, and popular tweets are more likely to contain positive emojis. What's more, the pattern held for every language tested. So far, it looks like we all prefer over .









