Event extraction is a complex task in information extraction, aimed at retrieving structured event details from text. With the advent of social media and related mobile applications, Internet-based communication technologies have grown rapidly, allowing people to exchange information on trending topics. Despite their importance, most of the relevant data is unavailable to humanitarian organizations in times of need, mostly because of various issues with data processing and data quality. ALBERT-enhanced graph attention network is proposed for extracting event information from text datasets. A pre-trained BERT model enhances the contextual information in the processed text. The proposed model is implemented into two phases: Event type extraction and event argument extraction. The proposed model is evaluated on the Maven dataset. According to the findings, the performance of the proposed model gives promising results on the Maven dataset.