Premier Analytics Consulting at the Southeast and Atlantic SAS® User Group (SEASUG) 2026 Conference

This hands-on workshop provides a practical introduction to building machine learning workflows that support modern AI systems using Python. Designed for data scientists, statisticians, programmers, machine learning engineers, researchers, analysts, and students, this workshop focuses on how to prepare data, train supervised learning models, tune model performance, and interpret predictive results for classification and regression tasks. Attendees will gain practical experience with Python’s open-source machine learning ecosystem, with an emphasis on organized, reproducible machine learning pipelines, model evaluation and selection, and fine-tuning strategies. Through guided examples and hands-on exercises, attendees will learn how machine learning models are developed as part of broader AI systems and how predictive models can be trained, evaluated, and interpreted within reliable analytical workflows. Emphasizing a model-driven approach, this workshop incorporates data preparation for machine learning pipelines, balancing model complexity with interpretability, and addressing common modeling challenges such as overfitting, underfitting, feature selection, and generalization to unseen data. Attendees will work within the scikit-learn ecosystem to build structured workflows for analytical and machine learning problems. Key topics covered in this workshop include: ➤ Data cleaning and exploratory data analysis (EDA) to uncover feature relationships Building comprehensive scikit-learn pipelines to clean data, engineer features, and prepare datasets for machine learning ➤ Training and interpreting supervised learning models, including LASSO regularization, decision trees, random forests, and gradient-boosted ensembles ➤ Hyperparameter tuning, search spaces, feature selection, and strategies for improving generalization to unseen data ➤ Model evaluation strategies, data partitioning techniques, and evaluation metrics for classification and regression ➤ Understanding the bias-variance tradeoff and model interpretability ➤ Connecting machine learning workflows to broader AI system design and applied analytics and decision-making This workshop helps attendees move beyond treating machine learning as a “black box” and instead develop a practical understanding of how predictive models are trained, tuned, evaluated, and interpreted in modern AI workflows. By the end of this workshop, participants will have built structured Python-based machine learning pipelines that can support reproducible analysis, predictive modeling, and applied AI system development. All registered attendees will receive non-redistributable PDF slides, fully documented Python notebooks, and workshop datasets so they can reproduce the analyses and continue practicing the methods after the tutorial session.