Discord Unveiled: A Comprehensive Dataset of Public Communication (2015-2024)

Yan Aquino, Pedro Bento, Arthur Buzelin, Lucas Dayrell, Samira Malaquias, Caio Santana, Victoria Estanislau, Pedro Dutenhefner, Guilherme H. G. Evangelista, Luisa G. Porfirio, Caio Souza Grossi, Pedro B. Rigueira, Virgilio Almeida, Gisele L. Pappa, Wagner Meira Jr

Read on arXiv

Type: paper
Venue: arXiv / UFMG
Year: 2026
Source: arxiv
Access: free
Language: en
Added: 2026-08-14T20:50:00Z
Verified: 2026-08-14T20:50:00Z

Summary

Largest claimed public Discord dataset: 2,052,206,308 messages from 4,735,057 users across 3,167 public Discovery servers (~10% of 31,673 servers listed as of 2024-11-17), spanning 2015-05-13 to 2024-12-17. Collected via Discord API; usernames pseudonymized (mimesis), IDs SHA-256 truncated to 12 chars. Per-server JSON plus servers_metadata. 17% of messages from bots. English-US dominates preferred_locale (1705 servers) with Spanish, French, Portuguese also present. Gaming still the top description keyword (~15%). Authors argue Discord is understudied vs Twitter/Reddit and useful for user-driven moderation research. Paper submitted to ICWSM 2025.

Keywords

discord · social-media · public-chat · moderation · bots · icwsm · zenodo · ufmg

Topics

computational social science, Discord, public chat dumps

Research notes

Primary: arxiv abs (cs.SI/cs.DB; Submitted to ICWSM 2025). Paper cites Zenodo DOI 10.5281/zenodo.14658505; that record 404 at 2026-08-14 check. HF paper page links unofficial mirrors fvdfs41/Discord-Unveiled and SaisExperiments/Discord-Unveiled-Compressed — not treated as official. Anonymized public Discovery servers only. Dataset not in datasets_local.csv. Discord posted abs.