Gulas, Christian

1 dataset credited to Gulas, Christian.

1 dataset.

SMIDGE Telegram Disinformation Dataset Analyzed

A Telegram (86 disinformation/conspiracy channels) dataset of 279,682 cleaned messages across 83 channels, collected by Federation of American Scientists via Telegram API. Each record contains message text (cleaned), channel name, timestamp, extracted URL, domain, and detected language per message. SMIDGE is a curated Telegram corpus spanning 2023-09-25 to 2024-09-24, focused on 83 channels selected for disinformation and conspiracy content, with 5,334 unique domains linked. Suitable for disinformation detection, conspiracy theory analysis, NLP text classification, content moderation, cross-platform information spread research.

279.7K records2026