The unprecedented COVID-19 pandemic exposed critical weaknesses in global health management, particularly in resource allocation and demand forecasting. This work introduces a transformative approach to enhance pandemic preparedness through real-time social media analysis. We present CoViNAR, a novel dataset of 14,000 annotated tweets categorized as "Need" or "Availability" of resources (NAR) during COVID-19. The proposed method combines BERTopic for context-aware filtering with advanced machine-learning techniques. The NAR classification model achieved an accuracy of 96.42%, and an F1-score of 96.43% using SVM with DistilBERT embeddings. We demonstrated the success of implementing the NAR classification model by doing a temporal analysis of tweets from the US, UK, and India for the duration of November 2019 to March 2023. The strong correlation between NAR tweet counts and COVID-19 case surges indicates the effectiveness of the proposed method, offering health authorities a powerful, proactive tool for resource management, that can positively transform crisis response strategies. The presented model successfully addresses a crucial gap in the design of early warning systems.