The document walks through a structured analysis pipeline to identify key predictors affecting concrete strength. It begins with loading a dataset, renaming and cleaning the data, and applying data normalization techniques. The study then develops multiple linear regression models, tests for model assumptions, and applies transformations (Box-Cox, square root) to improve model fit and interpretability. The analysis also includes influence diagnostics (Cook’s distance, leverage plots) and residual analysis to validate model robustness. Key variables such as Cement, BFS (Blast Furnace Slag), Water, Superplasticizer, Coarse Aggregate, Fine Aggregate, and Age are evaluated for their contribution to compressive strength. Box-Cox analysis suggests a square root transformation is optimal. EDA visuals (histograms, scatter plots, pair plots) highlight variable distributions, correlations, and skewness. The report concludes with detailed interpretation of variable effects on concrete strength, identifying Cement and Age as the strongest positive contributors, and Water as the most negatively correlated factor. It provides a summary of statistical findings, model summaries, and recommendations for predictive modeling improvements.

This project applies both K-Means and Hierarchical clustering on the Heart Failure dataset. The analysis includes data preprocessing, scaling, optimal cluster selection using the Elbow Method and Silhouette Score, and evaluation using both internal and external validation metrics.

Análisis estadístico espacial utilizando el índice de calidad del entorno