'Exploratory Data Analysis' (ED

EDA (pioneered by John Tukey) involves univariate analysis (distributions, histograms, box plots), bivariate analysis (correlation matrices, scatter plots), multivariate analysis, missing value analysis, outlier detection, skewness/kurtosis examination, and feature relationship exploration. EDA informs feature engineering, model selection, and data cleaning decisions. In banking, EDA on loan application data might reveal that income-to-EMI ratio is highly correlated with default which is a critical feature to include in the credit model. Tools used for EDA include Pandas Profiling, Seaborn, Matplotlib and Plotly.