data-summarizer

# Data Summarizer ## Quick Start Analyze datasets to produce statistical summaries, identify trends, and extract key insights. Support CSV, JSON, and tabular data formats. Present findings in clear tables and bullet points. ## Workflow 1. Load and inspect the dataset: format, row count, column types, missing values 2. Generate descriptive statistics: mean, median, min, max, std dev for numeric columns 3. Identify data quality issues: nulls, outliers, duplicates, inconsistent formats 4. Analyze distributions and trends for key metrics 5. Extract top insights and notable patterns 6. Present a structured summary with visualizable recommendations ## Examples **Example 1: Sales Data Summary** Input: "Analyze this Q1 sales data file" Output: ``` DATA SUMMARY ============ Dataset: Q1 Sales Data Records: 2,847 rows × 12 columns Period: 01/01/2026 – 03/31/2026 DATA QUALITY ------------- | Check | Result | |----------------|---------------------------| | Missing Values | 23 rows (0.8%) in 'region'| | Duplicates | 5 duplicate order IDs | | Outliers | 3 orders > $25,000 | DESCRIPTIVE STATISTICS ---------------------- | Metric | Value | |---------------|--------------------| | Total Revenue | $12.45M | | Avg Order | $4,372 | | Median Order | $2,850 | | Max Order | $580,000 | | Min Order | $125 | KEY INSIGHTS ------------ 1. Revenue grew 18% month-over-month (Jan→Feb→Mar) 2. Top 3 products account for 45% of total revenue 3. New York metro area leads with 38% of orders 4. Weekend orders are 35% lower than weekday average 5. Average order value increased 12% compared to Q4 2025 RECOMMENDED VISUALIZATIONS: - Line chart: monthly revenue trend - Bar chart: revenue by region - Pie chart: product category distribution ``` **Example 2: Quick Stats** Input: "Give me a quick summary of this CSV file" Output: A concise overview with row/column count, data types, missing values, and top-5 statistics. ## Tools - Use `Read` to load CSV, JSON, or data files - Use `Bash` to run data processing commands (sort, count, aggregate) - Use `Write` to save analysis reports - Use `Grep` to search for specific patterns in data files ## Error Handling - If file format is unsupported → suggest converting to CSV or JSON - If dataset is too large to process at once → sample or process in chunks - If columns have mixed data types → flag and suggest cleanup - If no numeric columns found → focus on categorical analysis (counts, modes, distributions) ## Connectors (Optional) This skill works standalone. When connected to external tools, it unlocks additional capabilities: | Connector | What it enables | |-----------|----------------| | ~~database | Query live databases directly to pull datasets for analysis | | ~~BI tool | Push summaries and insights into dashboards and automated reports | | ~~spreadsheet | Read data from linked spreadsheets and write summary tables back | | ~~data warehouse | Access large-scale datasets with full schema and partition metadata | ## Rules - Always start with data quality assessment before analysis - Report exact numbers — never round without stating the original value - Flag outliers but do not remove them without user confirmation - Currency defaults to USD; adjust when data indicates otherwise - Percentages should include the base number for context - Suggest appropriate chart types for key findings - Note any assumptions made during analysis ## Output Template ``` DATA SUMMARY ============ Dataset: [Name/Description] Records: [Rows] × [Columns] Period: [Date range if applicable] DATA QUALITY ------------- | Check | Result | |-------|--------| | [Check type] | [Finding] | DESCRIPTIVE STATISTICS ---------------------- | Metric | Value | |--------|-------| | [Stat] | [Value] | KEY INSIGHTS ------------ 1. [Insight with supporting data] RECOMMENDED VISUALIZATIONS: - [Chart type]: [What it shows] ``` ## Related Skills - **chart-generator** -- turn key findings into polished visualizations and dashboard components - **kpi-tracker** -- frame summary insights as trackable KPIs with targets and scorecards - **data-cleaner** -- fix data quality issues discovered during the profiling step