Russian and Ukrainian War-related Telegram Channels Dataset

This dataset provides a large-scale collection of Telegram channels and posts related to the ongoing Russo-Ukrainian war. It captures a wide range of war-related content, including news, commentary, propaganda, and public discussions, making it a valuable resource for researchers studying online information flows, disinformation, and digital discourse surrounding the conflict.

Dataset Structure

1. channels.csv

A CSV file containing metadata for 66,260 Telegram channels.

Columns:

• id: Unique identifier of the Telegram channel.

• title: Name of the Telegram channel.

• date: Channel creation date.

• fake: Binary indicator (true/false) showing whether the channel is flagged as fake.

• username: Channel’s username (if available).

• about: Channel description.

• pts: Telegram API persistent identifier for the latest event in the update sequence.

• participants_count: Number of subscribers at the time of data collection.

• pinned_msg_id: ID of the pinned message (if any).

• linked_chat_id: ID of the linked chat (if any).

2. leiden_clusters.csv

A CSV file containing clustering results of 3,074 channels, grouped into 8 clusters based on content and interaction patterns using the Leiden algorithm.

Columns (in addition to standard channel metadata):

• peer_id: Unique channel identifier.

• cluster: Assigned cluster number.

• war: Mentions of “war” in posts.

• svo: Mentions of “СВО” (Special Military Operation).

• na_ukraine: Mentions of the phrase “на Украине” (in Ukraine).

• v_ukraine: Mentions of the phrase “в Украине” (in Ukraine).

• oppo: Mentions of opposition-related terms (definition pending).

• slava: Mentions of “слава” (glory).

• sila: Mentions of “сила” (strength).

• posts_count: Total number of posts per channel.

• posts_count_wartime: Number of posts published during wartime.

• last_post: Timestamp of the latest post.

• ua_symbol: Mentions of Ukrainian symbols.

• ru_symbol: Mentions of Russian symbols.

• uk: Share of Ukrainian-language usage.

• ru: Share of Russian-language usage.

• antirus: Mentions expressing anti-Russian sentiment.

• antiukr: Mentions expressing anti-Ukrainian sentiment.

• war_engagement: Engagement score with war-related content.

• toxic: Toxicity score of channel content.

• score: Ideological leaning score (Pro-Russian vs Pro-Ukrainian).

• side: Channel alignment (Pro-Russian, Pro-Ukrainian, Neutral).

3. post_texts_part_{n}.csv

Multiple CSV files containing post texts, split for scalability.

• 79,572,856 posts from 4,687 channels.

• Coverage: 2015-09-22 (earliest channel creation date) to 2024-03-26 (most channels up to 2024-02-01).

Columns:

• id: Unique identifier of the post.

• peer_id: Channel ID.

• date: Post timestamp.

• message: Post text.

• views: Number of views.

• forwards: Number of forwards.

• edit_date: Timestamp of last edit (if applicable).

4. post_fwd.csv

A CSV file documenting 18,190,690 forwarded posts.

Columns:

• id: Unique identifier of the forwarded post.

• peer_id: ID of the channel that forwarded the post.

• fwd_from_channel_id: ID of the original source channel.

• fwd_from_channel_post: ID of the original post.