<p>Understanding population distribution is essential for urban planning and infrastructure development, especially in rapidly growing cities like Patna, India. Traditional methods often overlook spatial clustering and thresholding of population data, limiting their utility in high-resolution urban studies. This research addresses the gap by introducing a robust geospatial and statistical framework to identify population density clusters using remotely sensed raster data. The study utilized the World Population 2020 dataset and Google Earth Engine (GEE) for data extraction, with Patna defined using administrative boundaries from the GAUL dataset. Analysis revealed that Patna’s population density is highly skewed, with dense urban pockets and a peak value of over 13,700 persons/km². To segment high and low-density areas, Otsu’s thresholding method was applied, providing an optimal threshold value of 6802.24. Following this, k-means clustering was performed on pixels exceeding the threshold, and silhouette analysis was used to determine the best number of clusters. Results indicated that two clusters best represented the population distribution, with their centres located at specific geographic coordinates. The findings highlight areas of concentrated population, offering a practical approach for targeting interventions and resources. This method effectively combines image processing and statistical clustering, enhancing population mapping accuracy. The study provides a scalable model for urban demographic analysis. Future research could incorporate temporal datasets to monitor urban expansion and integrate socio-economic data for a more comprehensive understanding of population dynamics, helping policymakers and urban planners make data-driven decisions.</p>