Tabular Analyzer

Analyzes small tabular datasets (CSV, TSV, or Markdown tables) to compute objective descriptive metrics: row counts, distinct values, null percentages, min/max, and averages.

1. Purpose

Provide exact numerical facts about a dataset. "Descriptive analysis" means calculating observable statistical metrics without drawing unproven causal theories.

2. When to Use

Use this skill when:

• Analyzing test metrics, API latency logs, sales tables, or bug reports.
• Checking data quality and null rates in small tables (< 5,000 rows).
• User requests: "Analyze this table" or "Give me summary statistics".

3. When Not to Use

• Do not use to validate CSV schema rules; use csv-validator.
• Do not use to extract isolated key-value pairs; use field-extractor.

4. Required Inputs

• table_data: CSV text or markdown table containing columns and rows.
• numeric_columns: Target columns to compute mathematical statistics on.

5. Missing-Input Behavior

1. If table_data is empty: stop and ask: "Please provide the table or data rows."
2. If table_data has fewer than 2 rows: state "Insufficient sample size (fewer than 2 rows)."
3. Never invent synthetic data to pad the table.

6. Procedure

Follow these numbered actions:

1. Action 1: Calculate Dimensions
   • Count total rows and list all column names.
2. Action 2: Calculate Data Quality Metrics
   • For each column, count null/empty cells and compute null percentage.
3. Action 3: Calculate Numeric Metrics
   • For numeric columns: calculate Min, Max, and Mean (Sum / Count).
4. Action 4: Calculate Categorical Distributions
   • For string columns with <= 5 unique values: list frequency count per category.
5. Action 5: Emit Analysis Report
   • Format using Section 8.

7. Decision Rules

• If non-numeric strings appear in a numeric column: count them as Invalid / NaN and report the error rate.
• Never state "this trend proves X" without statistical significance testing; describe only observable facts.

8. Output Format

Emit the analysis using this template:

### Tabular Analysis Report

- **Total Rows**: [Count]
- **Total Columns**: [Count]

#### Column Quality Metrics
| Column | Data Type | Null Count | Null % | Unique Values |
| :--- | :--- | :--- | :--- | :--- |
| `[col1]` | Integer | 0 | 0.0% | [Count] |
| `[col2]` | String | 2 | 4.0% | [Count] |

#### Numerical Summary
| Column | Min | Max | Mean | Notes |
| :--- | :--- | :--- | :--- | :--- |
| `[num_col]` | [Min] | [Max] | [Mean] | [Clean data] |

#### Observational Summary
- [Objective observation 1 with exact numbers]
- [Objective observation 2]

9. Verification Checklist

• Row and column counts match data exactly.
• Means are calculated accurately (sum / count).
• No speculative or unproven inferences were added.
• Null percentages are correctly computed.

10. Common Failures and Recovery

11. Stop or Escalation Conditions

Stop and ask the user when:

• Dataset contains corrupted binary or unaligned columns.

12. Examples

Example 1: Successful Execution

Input:

Table:
id,latency_ms,status
1,120,200
2,150,200
3,300,500
4,110,200

Output:

### Tabular Analysis Report

- **Total Rows**: 4
- **Total Columns**: 3

#### Column Quality Metrics
| Column | Data Type | Null Count | Null % | Unique Values |
| :--- | :--- | :--- | :--- | :--- |
| `id` | Integer | 0 | 0.0% | 4 |
| `latency_ms` | Integer | 0 | 0.0% | 4 |
| `status` | Integer | 0 | 0.0% | 2 |

#### Numerical Summary
| Column | Min | Max | Mean | Notes |
| :--- | :--- | :--- | :--- | :--- |
| `latency_ms` | 110 | 300 | 170.0 | Outlier at 300 ms on row 3 |

#### Observational Summary
- 3 out of 4 requests (75%) returned HTTP 200.
- 1 request (25%) returned HTTP 500 with the highest observed latency (300 ms).

Example 2: Ambiguous or Missing Input

Input:

Table:
(Only header row: id,name)

Output:

### Tabular Analysis Report: Insufficient Data

- **Total Rows**: 0 (Header only)
- **Total Columns**: 2

Cannot compute statistics on an empty table. Please provide at least 2 rows of data.

Example 3: Expected Final Output

Output:

### Tabular Analysis Report

- **Total Rows**: 3
- **Total Columns**: 2

#### Numerical Summary
| Column | Min | Max | Mean | Notes |
| :--- | :--- | :--- | :--- | :--- |
| `score` | 80 | 100 | 90.0 | All scores positive |

#### Observational Summary
- Mean score is 90.0 across 3 records.