exploratory-data-analysis

Multi-stage EDA pipeline (6 major stages with sub-stages) for turning raw business data into validated, confidence-labeled analytical notebooks. Use this skill whenever the user asks for exploratory data analysis, data investigation, business analytics, dataset profiling, hypothesis testing on tabular data, multi-table analysis, customer segmentation, market basket analysis, or any multi-step analysis workflow — even if they don't say "EDA" explicitly. Trigger on: "analyze this data", "what can we learn from this CSV", "profile this dataset", "find insights in this data", "help me understand this data", "run analytics on", "build an analysis notebook", inventory analysis, sales analysis, churn analysis, cohort analysis, any request involving CSV/Excel/Parquet data where the user wants business-actionable findings, any request involving multiple related tables, or any request where the user provides a dataset and asks open-ended questions about it. Also trigger when the user references a Kaggle dataset, competition data, or any structured analytical project with more than one data file.