Searching Emojis With Casual Japanese Keywords — Why Unicode CLDR's ja Annotations Aren't Enough

Short summary

Dev curated 107 emojis with casual Japanese search tags, exposing a gap between Unicode CLDR's accessibility-focused annotations and how users actually search (e.g., 'ぴえん' for pleading face). Built a linear-scan search with five-tier scoring that prioritizes exact tag matches over substring hits, handling NFKC normalization and mixed-kana/kanji input. Includes working demo and GitHub implementation.

• •CLDR's official emoji annotations optimize for accessibility captions, not user search behavior—critical gap for international products
• •Hand-curated casual Japanese tags reveal actual user language ('ぴえん' slang, mixed kana/kanji, English loanwords like 'lol')
• •Scoring algorithm with configurable weights ensures tag matches outrank coincidental substring hits in emoji names

Generated with AI, which can make mistakes.