Abstract
This study develops an automated framework for land use classification using Landsat 8 imagery processed on Google Earth Engine (GEE), directly applied to smart grid planning. The methodology combines spectral indices (NDVI, NDWI, NDBI) with optical bands(B2-B7) and applies k-means clustering for unsupervised classification into five land cover types: urban, agricultural, forest, water, and barren areas. Applied to Astana, Kazakhstan [71.19-71.61°E, 51.025-51.175°N], the classification reveals agricultural dominance (38.5%), significant urban coverage (31.2%), and substantial forest areas (23.7%). An interactive visualization interface with dynamic inspection tools enhances stakeholder accessibility. Beyond conventional environmental monitoring, classified outputs directly inform smart grid development: urban zones indicate high-load demand areas for electric vehicle infrastructure; agricultural land represents biomass energy potential; water bodies support hydropower assessment; barren areas identify solar farm sites. By quantitatively linking land cover data with energy metrics, this framework provides actionable inputs for renewable integration, demand forecasting, and sustainable urban energy planning. Future research will integrate classified datasets with smart grid simulation platforms for distributed generation optimization.