I have a dataset that blends numerical fields with several categorical factors and I need a clean, defensible multiple-linear-regression model that can reliably predict my target variable. My immediate priorities are: • Building the baseline model, encoding categorical predictors appropriately and confirming the functional form of each term. • Running full model-adequacy checks—residual diagnostics, multicollinearity assessment, influence analysis, and tests for normality and constant variance. • Proposing and applying any necessary variable transformations (log, Box-Cox, polynomial, interaction terms, or recoding of factor levels) so that the underlying assumptions are met and predictive performance improves. Please document each step clearly in code (R, Python, or another mainstream statistics tool you prefer) and produce a concise report that explains what was tried, what was retained or discarded, and why. Final deliverables are: 1. Script or notebook with reproducible analysis and inline comments. 2. PDF or markdown summary of findings, diagnostics, and recommended next actions. 3. Tidy dataset (or data-processing script) reflecting any transformations applied. Acceptance criterion: the final model must pass all standard diagnostic checks without violating key assumptions, and it should outperform the untreated baseline on relevant error metrics (e.g., RMSE, MAE).