Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Longitudinal, Multinational, and Multilingual Corpus of News Coverage of the Russo-Ukrainian War

Dikshya Mohanty, Taisiia Sabadyn, Jelwin Rodrigues, Chenlu Wang, Abhishek Kalugade, Ritwik Banerjee

TL;DR

DNIPRO addresses the need for a longitudinal, multinational, multilingual corpus of war coverage by introducing 246,229 news articles from 11 outlets across 5 countries and 3 languages, spanning Feb 2022 to Aug 2024. The paper describes a multi-stage pipeline including data acquisition, translation to English, and rich annotations (NER, sentiment, stance, topical framing) with rigorous human evaluation, stored in Parquet for efficient analysis. Use-case experiments demonstrate the dataset's utility for examining thematic coverage, framing, stance, and narrative divergence, revealing polarized narratives across geopolitical actors. DNIPRO thus provides a foundational resource for cross-cultural discourse analysis, computational journalism, and investigations of information warfare dynamics in global information ecosystems.

Abstract

We introduce DNIPRO, a novel longitudinal corpus of 246K news articles documenting the Russo-Ukrainian war from Feb 2022 to Aug 2024, spanning eleven media outlets across five nation states (Russia, Ukraine, U.S., U.K., and China) and three languages (English, Russian, and Mandarin Chinese). This multilingual resource features consistent and comprehensive metadata, and multiple types of annotation with rigorous human evaluations for downstream tasks relevant to systematic transnational analyses of contentious wartime discourse. DNIPRO's distinctive value lies in its inclusion of competing geopolitical perspectives, making it uniquely suited for studying narrative divergence, media framing, and information warfare. To demonstrate its utility, we include use case experiments using stance detection, sentiment analysis, topical framing, and contradiction analysis of major conflict events within the larger war. Our explorations reveal how outlets construct competing realities, with coverage exhibiting polarized interpretations that reflect geopolitical interests. Beyond supporting computational journalism research, DNIPRO provides a foundational resource for understanding how conflicting narratives emerge and evolve across global information ecosystems.

A Longitudinal, Multinational, and Multilingual Corpus of News Coverage of the Russo-Ukrainian War

TL;DR

DNIPRO addresses the need for a longitudinal, multinational, multilingual corpus of war coverage by introducing 246,229 news articles from 11 outlets across 5 countries and 3 languages, spanning Feb 2022 to Aug 2024. The paper describes a multi-stage pipeline including data acquisition, translation to English, and rich annotations (NER, sentiment, stance, topical framing) with rigorous human evaluation, stored in Parquet for efficient analysis. Use-case experiments demonstrate the dataset's utility for examining thematic coverage, framing, stance, and narrative divergence, revealing polarized narratives across geopolitical actors. DNIPRO thus provides a foundational resource for cross-cultural discourse analysis, computational journalism, and investigations of information warfare dynamics in global information ecosystems.

Abstract

We introduce DNIPRO, a novel longitudinal corpus of 246K news articles documenting the Russo-Ukrainian war from Feb 2022 to Aug 2024, spanning eleven media outlets across five nation states (Russia, Ukraine, U.S., U.K., and China) and three languages (English, Russian, and Mandarin Chinese). This multilingual resource features consistent and comprehensive metadata, and multiple types of annotation with rigorous human evaluations for downstream tasks relevant to systematic transnational analyses of contentious wartime discourse. DNIPRO's distinctive value lies in its inclusion of competing geopolitical perspectives, making it uniquely suited for studying narrative divergence, media framing, and information warfare. To demonstrate its utility, we include use case experiments using stance detection, sentiment analysis, topical framing, and contradiction analysis of major conflict events within the larger war. Our explorations reveal how outlets construct competing realities, with coverage exhibiting polarized interpretations that reflect geopolitical interests. Beyond supporting computational journalism research, DNIPRO provides a foundational resource for understanding how conflicting narratives emerge and evolve across global information ecosystems.
Paper Structure (21 sections, 3 figures, 7 tables)

This paper contains 21 sections, 3 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. The dnipro Corpus
  4. Data Acquisition
  5. Metadata Schema and Statistics
  6. Data Records
  7. Data Annotations
  8. Text Processing for Annotation
  9. Machine Translation
  10. Annotation Methodology
  11. Human Evaluation
  12. Topical Frames:
  13. Stance Annotations:
  14. Use Case Experiments
  15. Thematic Analysis of Conflict Events:
  16. ...and 6 more sections

Figures (3)

  • Figure 1: Composition of dnipro, across all sources, languages (English, Russian$\dagger$, and Mandarin$\ddagger$), and national affiliations: Censor, Kyiv Independent, and European Pravda from Ukraine; Sputnik, TASS, and Izvestia$\dagger$ from Russia; CNN from USA; Financial Times from UK; and China Daily, Global Times, and Xinhua$\ddagger$ from China. Above, we see the distribution of articles (left) and their average size (right).
  • Figure 2: Topical framing by source, with % share of articles for each label (confidence $>$ 0.5). Source abbr.: Financial Times (FT), Kyiv Post (KP), Sputnik (SP), European Pravda (EP), Global Times (GT), China Daily (CD), Xinhua (XH). Labels: (Sec)urity, (Pol)itics, (Econ)omy, (Welf)are, (Ident)ity, (Moral)ity, and (Pub)lic (Op)inion.
  • Figure 3: Daily average of stance scores towards Ukraine, grouped by media sources from five nations. The first 15 days of the war (left), contrasted with the last 15 days collected for dnipro (right).