feat: complete data preparation pipeline with side-by-side outlier visuals and processed dataset

refactor: clean duplicate plots, update preprocessing script, and add outlier treatment visuals

Final Submission: Integrated Streamlit app, Dockerfile, and CI/CD workflow