q_descriptive-analysis_skill

Overview

This skill performs comprehensive descriptive analysis of tabular datasets, producing grouped statistics, frequency distributions, entity extraction from text fields, temporal trend tables, and publication-ready markdown summaries. It outputs CSV tables for validation and an MD summary for interpretation, making results easy to review or include in reports.

How this skill works

I inspect the dataset, ask targeted questions about objectives, grouping variables, continuous measures, text fields, temporal columns, and output preferences, then prepare derived variables (tiers, time periods). The skill computes overall and stratified descriptive statistics, builds frequency tables for categorical fields, extracts and normalizes entities from text, and aggregates metrics over time. Final outputs are structured CSVs and a markdown summary linking the key tables and highlighting top values and temporal patterns.

When to use it

• Exploratory data analysis for Excel or CSV datasets before modeling or hypothesis testing
• Preparing publication-ready descriptive tables for academic manuscripts or reports
• Summarizing engagement metrics, SDT scores, or any continuous/categorical measures by subgroup
• Extracting named elements from free-text fields (timestamps, named entities, annotations)
• Analyzing temporal dynamics and creating period-level trend tables for presentations

Best practices

• Define research objective and grouping variables up front to avoid redundant outputs
• Limit top-N displays for large cardinality categorical fields (top 5–10) and provide full CSVs
• Standardize date/time formats before temporal aggregation to ensure correct bins
• Provide or confirm custom classification thresholds (tiers) for consistent stratification
• Validate entity extraction patterns on a small sample and adjust regex rules if needed

Example use cases

• Generate overall and by-group descriptives for video engagement metrics (views, likes, watch time) and export CSVs for each grouping variable
• Extract speaker or event names from timestamped transcript notes, produce entity frequency tables, and include top entities in the MD summary
• Create monthly trend tables for enrollment or usage metrics and annotate milestones for interpretation
• Produce continuous variable summaries (mean, SD, median, IQR) overall and stratified by predefined tiers for an academic supplement
• Deliver a folder of CSVs plus a DESCRIPTIVE_SUMMARY.md that reviewers can use to verify reported numbers