[Read&Explore] Studying Cultural Differences in Emoji Usage across the East and the West
This paper studuies posts from the East (China and Japan) and the West (the US, UK, and Canada) to investigate their differences in Emoji Usage.
Data Collection and Preparation
- The authors collected posts from Weibo and Twitter in 2014 as user data. Weibo for China, and Twitter for Japan and the West.
- The coordinates on Twitter and self-reported location on Weibo are used to locate users' countries.
- They only keeped posts written in monolingual language.
- The authors first tokenized the text of each post, and trained a Word2Vec model to learn an embedding for each token (including emojis).
- They generated the embedding for each LIWC category by averaging the word embeddings contained in it.
- They computed the frequency of emojis in each country.
- They computed a pair-wise similarity for emojis among countries using cosine similarity of embeddings to see the general similarity between two countries.
- They computed a pair-wise similarity for emojis and LIWC categories to see which emojis was the most frequent in each category for each country.
The highlights in my POV:
- The use of emojis :weary: :ok_hand: :smirk: :raised_hands: :unamused: :100: is more widespread in the West than in the East.
- Conversely, the use of emojis :pray: :scream: :muscle: :thumbsup: :flushed: :sparkles: :stuck_out_tongue_closed_eyes: is more common in the East than in the West.
- Emojis are used similarly across cultures when conveying universal themes such as "Ingest," "Death," "Anger," "Money," and "Home."
- However, there are differences in the use of emojis when expressing more nuanced issues like "Insight," "Number," "Time," and "Friend."