<p>Flood data collected through citizen science initiatives in Balikpapan City requires further transformation, analysis, and processing to become valuable information for decision-making and policy development. The study’s goal is to evaluate Balikpapan City’s flood severity data for 2024 and 2025 by clustering the recorded flood parameters and severity labelling the clusters of flood data. The flood data that have been collected, which include the coordinates’ location, flood height, and duration in the study area, can be leveraged using clustering or unsupervised learning. Methods that were employed consisted of K-means clustering and K-medoids clustering that also known as partitioning around medoids or PAM, to extract meaningful insights. The clustering results are assessed to identify the appropriate clusters for evaluating flood severity levels in the research area. Since the normal flood level in publications is based on vulnerability, the clustering-based severity levelling is the novel proposed flood levelling or classification. The labelled clustered data will significantly impact flood disaster management planning, particularly in prioritizing mitigation strategies. The result of the study will also trigger new research in further research dealing with flood severity, which is less than other flood-related papers.</p>