Emoji Stats

The Unicode emoji dataset is a richly structured corpus with multiple analytical dimensions. At the top level, emoji are grouped into nine categories in emoji-test.txt, but these groups are merely organizational conventions — the normative metadata lives in emoji-data.txt, which assigns each codepoint properties like Emoji, Emoji_Presentation, Emoji_Modifier, Emoji_Modifier_Base, and Emoji_Component. These properties govern how emoji behave in text processing, input methods, and rendering engines.

From a growth perspective, the emoji set has expanded from the original 722 emoji in Unicode 6.0 (2010) — largely imported from Japanese carrier encodings — to nearly 4,000 in Emoji 16.0. Each annual release cycle involves formal proposals submitted to the Unicode Emoji Subcommittee, review by the Unicode Technical Committee, and a beta period for public comment. Proposals are evaluated against criteria including expected usage frequency, distinctiveness from existing emoji, and breadth of appeal across cultures and geographies. Many proposed emoji are rejected on the grounds of excessive similarity to existing characters or insufficient evidence of widespread demand.

Sequence emoji — including skin tone modifier sequences, ZWJ sequences, keycap sequences, and flag sequences — account for a significant portion of the emoji catalog but share relatively few base codepoints. This architectural efficiency means that adding a new base emoji character can enable many derived sequences. Statistical analysis of the emoji dataset reveals patterns in design: the 'People & Body' group is the largest due to the combinatorial explosion of skin tone and gender variants, while 'Flags' forms a distinct cluster defined almost entirely by sequence mechanics rather than individual glyph design.