Exploratory Data Analysis by Building (two projects in parallel)

OBJECTIVE: do a professional exploratory data analysis (EDA) on two real projects at the same time — German Credit (classification, plus our own customer personas and risk tiers) and Spain Electricity Demand (regression, time series) — from framing the problem to a clean, training-ready dataset. WHAT WE COVER: • Workspace, helper toolkit and a one-page problem brief (costs of errors, metric, hypotheses, leakage watch-list) • Loading, joining and integrity checks; data quality and missing values • Target analysis, leakage and fairness checks • Univariate analysis of numerical and categorical features • Relationships between features: correlation, multicollinearity, mutual information, interactions • Outliers and anomalies; time-based EDA (trend, seasonality, lags, drift) • Feature engineering plan and leakage-safe preparation • A small segmentation: personas and risk tiers • Feature shortlist, EDA report and a training-ready dataset hand-off AFTER THIS MODULE YOU CAN: • Run a complete EDA on any tabular dataset and explain it to a business audience • Spot data problems and leakage before modelling • Hand over a clean, reproducible dataset and pipeline for machine learning