This notebook tells the “Cities of Tomorrow” story by taking the public Sustainable Urban Planning & Landscape dataset from raw CSV to production-ready insight. After enforcing clean schemas and confirming spotless data quality, it engineers a new green-space-per-capita proxy, visualizes the city ecosystem through correlations, regressions, outlier scans, and K-Means clusters, and then trains a Random Forest—tuned with 5-fold cross-validation and benchmarked against Gradient Boosted Trees—to predict urban sustainability scores. The winning model is logged, versioned, and packaged via MLflow, residuals are audited for bias and operational follow-ups, and the narrative closes with actionable recommendations for city planners and guidance on how the Fabric-and-Azure stack can operationalize sustainable, livable metropolis strategies.
https%3A%2F%2Fgithub.com%2FMiftahRizky%2FContest%2Fblob%2Fmain%2FNov2025_MiftahRizky_CitiesOfTomorrow.ipynb
