In the context of Local Digital Twins (LDTs), cities face the challenge of moving beyond static models and rule-based systems to leverage pattern recognition and adaptive decision-making capabilities. This is crucial for improving efficiency, resilience, and decision-making processes without unnecessarily increasing system complexity, especially for resource-constrained municipalities. To address this, data processing and machine learning (ML) are employed to enhance the functionality of LDTs. This involves transitioning from basic awareness to utilizing applied ML techniques like linear regression, decision trees, and support vector machines (SVMs) that support forecasting, anomaly detection, and categorization tasks efficiently and transparently. This unit focuses on teaching ML as a practical tool through a structured learning approach. Participants will engage with content that emphasizes data quality and preparation, practical ML methods tailored to specific city challenges, and critical reflection on model choice, ensuring methods support resilience and transparency. Trainers will guide learners to think like city planners and match the right tool to specific scenarios.