I have an Excel workbook containing 58 input features and six width-related targets. I need a solid regression pipeline that can learn the relationships and reliably predict all six outputs for new records. What I expect: • Cleanly written Python code (Pandas, scikit-learn or similar) that reads the Excel data, handles basic preprocessing, and splits it into train / validation sets. • Exploration of at least one baseline model plus one or two stronger algorithms—feel free to test linear methods, tree-based ensembles, or any other multi-output regressors you trust. • Hyper-parameter tuning and cross-validation so we avoid overfitting. • Clear evaluation using R² and MAE (or other sensible metrics) for every target, followed by a concise report explaining which model you chose and why. • A saved model file and a short notebook or script that shows how to load it and generate predictions on unseen data. The dataset is ready to hand over immediately, so you can dive straight into modelling without any collection work.