Datasetpublic

bruhwalkk/economic-telegram-news-corpus-2025

Economic Telegram News Corpus 2025 A corpus of 31,292 Russian-language economic news posts collected from 7 major Telegram channels, spanning January 2024 to September 2025. The dataset supports research on economic narrative detection, topic classification, and information diffusion in social media. Associated Paper Going Viral: LLM-Based Modeling of Economic Narratives Dataset Description The raw collection contains 123,273 posts. The economic… See the full description on the dataset page: https://huggingface.co/datasets/bruhwalkk/economic-telegram-news-corpus-2025.

sourceHugging Facecc-by-nc-4.0updated 5mo agoView on Hugging Face

1likes102downloads

Dataset Card

Economic Telegram News Corpus 2025

A corpus of 31,292 Russian-language economic news posts collected from 7 major Telegram channels, spanning January 2024 to September 2025. The dataset supports research on economic narrative detection, topic classification, and information diffusion in social media.

Associated Paper

Going Viral: LLM-Based Modeling of Economic Narratives

Dataset Description

The raw collection contains 123,273 posts. The economic corpus was constructed by:

1.Removing duplicates and near-duplicates
2.Excluding non-news content
3.Selecting posts assigned to economic topics via an LLM-based classifier (~90% accuracy on the Golden Set)

Virality Score

Each post includes a composite virality score (viral_final) computed over a 3-day window after publication:

viral_final = 0.45 * viral_static + 0.20 * viral_dynamic + 0.35 * viral_ml

Columns

Column	Type	Description
`message_id`	string (UUID)	Unique message identifier
`id_channel`	int	Channel ID (1–7)
`message`	string	Full post text (Russian)
`viral_final`	float	Composite virality score [0, 1]
`is_economic`	bool	Economic post flag
`economic_topic`	string	LLM-assigned economic topic (9 categories)
`topic_confidence`	float	Topic classification confidence
`channel_name`	string	Telegram channel name
`channel_w`	float	Channel weight
`message_vector`	string	PostgreSQL tsvector (full-text search index)
`subscribers`	int	Channel subscriber count
`date`	datetime	Exact publication timestamp (UTC)
`date_day`	datetime	Publication date (day-level, UTC)

Channels (7)

Channel	Posts
Forbes Russia	—
Блумберг	—
РИА Новости	—
Экономика	—
Раньше всех. Ну почти	—
Банки, деньги, два офшора	—
Сигналы РЦБ	—

Topics (9)

Topic	Count
Государственная экономическая политика	9,040
Корпоративные финансы	4,572
Макроэкономика	3,978
Санкции и геополитика	3,582
Рынки капитала	3,169
Сырьевые рынки	2,118
Международная торговля	1,825
Другое	1,608
Валютный рынок	1,400

Usage

python

from datasets import load_dataset

ds = load_dataset("bruhwalkk/economic-telegram-news-corpus-2025")

Citation

If you use this dataset, please cite the associated paper:

bibtex

@article{economic_narratives_2025,
  title={Going Viral: LLM-Based Modeling of Economic Narratives},
  journal={Записки научных семинаров ПОМИ},
  year={2025}
}

License

CC-BY-NC-4.0