Predicting Urban Area Expansion via Supervised Learning on Remote Sensing Data: A Practice on Tehran's Metropolitan Area

This research project employs satellite imagery analysis to model land cover change in Tehran's metropolitan area, with a specific focus on urban and built-up land from 2001 to 2023. Utilizing the MCD12Q1.061 MODIS Land Cover Type dataset, images were processed using Google Earth Engine (GEE), a cloud-based platform for geospatial data analysis, to efficiently handle the vast amount of data. This approach enabled the extraction of relevant information related to urban expansion, facilitating a comprehensive analysis of the dynamic urban landscape. The core analysis involved calculating pixel coverage representing urban and built-up areas and developing a linear regression model in Google Colab, leveraging the versatility of Python for data analysis and model development. The model established a strong correlation between time and urban expansion, achieving an R-squared value of 0.95, indicating that 95% of the variability in urban land area is explained by the linear trend over time. This robust fit suggests a consistent and predictable pattern of urban sprawl in Tehran. While a Mean Squared Error (MSE) of 7.36 suggests some deviations in predictions, the model's overall reliability is affirmed by its strong R-squared value, making it a valuable tool for forecasting future urban growth. The implications of this research extend beyond mere academic inquiry, holding significant relevance for urban development policies and sustainability initiatives. The quantitative insights generated by the study offer valuable data for urban planners seeking to make informed decisions regarding infrastructure development, resource management, and environmental sustainability, emphasizing the need for data-driven methodologies.