⚙️ Introduction

The project aimed to apply predictive modelling to clinical datasets, identifying patterns that forecast patient progress during rehabilitation and treatment.

🎯 Objectives

• Build predictive models for clinical outcomes
• Identify key predictors influencing recovery
• Compare classical statistical models with modern ML approaches
• Visualize predictive performance for interpretability

🧩 Methods

• Data wrangling and feature engineering in pandas and tidyverse
• Model training with logistic regression, random forests, and SVM
• Cross-validation and ROC/AUC evaluation
• Visualization of variable importance and partial dependence

💡 Outcome

Delivered a reproducible workflow for predicting patient outcomes, integrating data cleaning, modelling, and visualization. The project demonstrated the clinical potential of data-driven decision support.