Dataset and R Script for: Differentiated Communication Strategies in Intermestic Crises

Abstract Purpose: This study analyzes the differentiated communication strategies adopted by the national and local governments in Colombia during the 2025 intermestic crisis triggered by the Trump administration’s deportation and tariff policies. It explores how this episode exposed internal political divisions and empowers local leaders. Methodology: A quantitative design combining natural language processing and text mining was applied to 3,867 posts on X published during the first 100 days of the Trump administration. The dataset comprised 10 accounts (President, Presidency, and mayoral offices of Bogotá, Medellín, Cali, and Barranquilla). Analysis included engagement metrics, lexical frequency analysis, lexicon-based sentiment and emotion analysis using the NRC Emotion Lexicon. Results: The findings reveal a clear functional division in communication strategies. The national government prioritized an ideological and confrontational narrative representing 62% of its agenda and characterized by high emotional volatility and abstract language. Local governments adopted a more pragmatic approach, devoting 24%–60% of their agendas to governance and security issues. Strong personalization was identified with personal accounts generating up to 137 times more engagement than institutional accounts. Trust emerged as the dominant emotion. Conclusions: This study offers a multilevel perspective on political communication during an intermestic crisis, demonstrating how subnational actors use digital paradiplomacy to protect local interests. It also validates the integration of international relations theory with computational sentiment analysis as a governance tool. Keywords: political communication; local government; intermestic crisis; social media; sentiment analysis; diplomatic relations