TriEmojiMap is an interactive visualization tool designed to explore emoji semantics from three distinct AI-driven perspectives: semantic meaning (text), visual appearance (image), and contextual usage (language model embeddings).
Focusing on 1,810 emojis frequently used on social media, I created an interactive scrollytelling interface allowing intuitive exploration of emoji similarities and differences. Users can seamlessly transition between semantic, visual, and language-model layers, observing how emojis shift positions across different embeddings through animated interactions.
A dynamic slider enables users to identify emojis with significant discrepancies between visual appearance and actual usage, highlighting potential ambiguities. Additionally, a robust search feature lets users quickly find emojis by keyword or symbol, and clicking an emoji reveals a detailed information panel providing rich semantic descriptions, visual characteristics, and real-world usage contexts.
TriEmojiMap aims to provide an engaging, user-friendly way to explore and understand emoji meanings—recognizing that ultimately, each person has their own favorite emojis and unique ways of using them.
