A machine learning dissection of Nigeria’s ‘Renewed Hope Agenda’: Sentiment Dynamics and Thematic Salience in Public Discourse (2023–2025)

Adeyemi, Tewogbade Shakir (2025) A machine learning dissection of Nigeria’s ‘Renewed Hope Agenda’: Sentiment Dynamics and Thematic Salience in Public Discourse (2023–2025). International Journal of Science and Research Archive, 15 (3). pp. 1337-1348. ISSN 2582-8185

This research offers a computational assessment of Nigeria’s “Renewed Hope Agenda” under President Tinubu (May 2023–April 2025), probing whether its narrative framing has shifted public sentiment and policy‐issue salience in national newspapers. A corpus of 24,450 articles from the two major newspapers were collected and filtered for relevant keywords and minimum length. Lexicon‐based sentiment analysis (VADER) quantified monthly mean compound scores, revealing an overall positive tone (0.62–0.78), with notable dips following subsidy removal and inflationary events, and rebounds linked to communication recalibrations. Similarly, topic modelling using BERTopic with “paraphrase-MiniLM-L6-v2” embeddings identified 318 thematic clusters. Prominent topics included fuel subsidy removal, VAT reform, palliative measures, and digital transformation, while structural concerns like inflation, poverty, and insecurity persisted. Integration of NRC emotion lexicon via NRCLex enriched these themes with affective profiles, uncovering elevated anger and fear around economic policies and trust associated with governance transparency initiatives. Findings suggest that while “Renewed Hope” rhetoric maintains discursive traction, its legitimacy is contingent upon tangible socioeconomic outcomes. Methodologically, this research demonstrates the utility of combining sentiment analysis and embedding-based topic modelling for longitudinal policy discourse monitoring.