What problem does it solve? When you encounter a new table or data file, you often don't know its shape, quality, or which dimensions and metrics are worth analyzing. This Skill generates a comprehensive data profile so you can understand a dataset before writing analysis queries. ## Core Features & Use Cases - Automated Data Profiling: Computes row counts, null rates, cardinality, distributions, percentiles, and date ranges for every column, grouped by type (dimensions, metrics, dates, IDs). - Data Quality Detection: Flags high null rates, placeholder values, duplicates, inconsistent formats, impossible values, and suspicious distributions using a structured quality assessment framework. - Analysis Recommendations: Identifies the best dimension and metric columns, foreign key candidates, hierarchies, and suggests 3-5 concrete follow-up analyses. - Use Case: You connect a data warehouse and point the Skill at an unfamiliar orders table. It returns a profile showing 2.3M rows, flags a discount column with 40% nulls and negative amounts, and recommends a revenue trend analysis by region. ## Quick Start Ask the assistant to profile the table or uploaded file you want to understand, for example by saying: run explore-data on the orders table and summarize its quality issues.
