LiveClawBench

Validate tabular data against configurable business rules — range checks, cross-column constraints, referential integrity against lookup tables, and temporal consistency.

Load and parse CSV files into structured data frames with configurable encoding, delimiter, and date parsing options.

Export pandas DataFrames or processed datasets to well-formatted CSV files with configurable encoding, delimiter, quoting, and column ordering.

Aggregate tabular data by one or more dimensions with configurable metrics — sum, mean, count, min, max, median. Supports multi-level grouping and pivot table generation.

Comprehensive data cleaning toolkit — handles missing values, duplicate removal, type coercion, whitespace trimming, and outlier detection in a single pass over the dataset.

Identify and remove duplicate records from tabular datasets using exact-match, fuzzy-match, or composite-key-based deduplication strategies.

Universal data ingestion service that loads data from CSV, Excel, JSON, and Parquet sources into a unified DataFrame format with automatic format detection.

Read and parse Excel workbooks (.xlsx/.xls) with support for multiple sheets, named ranges, and automatic type inference.

Calculate derived columns and financial metrics from transaction data — revenue, profit margin, time-based features, moving averages, and growth rates.

Export datasets to multiple output formats in one invocation — CSV, JSON, Parquet, and Excel. Supports per-format configuration, batch export of multiple DataFrames, and ZIP packaging.

Detect and handle missing/null values in tabular datasets using configurable strategies — drop, fill-constant, fill-forward, fill-mean, or fill-median.

Normalize and scale numeric columns using min-max scaling, z-score standardization, or robust scaling. Supports saving and applying scaler parameters for consistent train/test normalization.

Validate tabular data against a predefined JSON schema — checks column presence, data types, nullable constraints, and unique constraints.

End-to-end ETL pipeline for retail sales data — ingests multi-format transaction files, cleans and validates data, computes derived financial metrics, aggregates across dimensions, and exports results with summary reports.

Generate comprehensive summary statistics and KPI reports from sales/transaction data — includes total revenue, top products, regional breakdown, channel performance, and time-series trends.

Aggregate data by multiple dimensions (time, product, region, channel) with configurable metrics, pivot tables, and generate KPI summary reports with top products, trends, and cross-dimensional breakdowns.

Comprehensive data cleaning — handles missing/null values (drop, fill-mean, fill-median, fill-constant), deduplication (exact and fuzzy), type coercion, whitespace trimming, and outlier detection in a single pass.

Export datasets to multiple formats — CSV (configurable delimiter, encoding, quoting), JSON (records/nested/lines), Parquet (compressed), and Excel (.xlsx with formatting). Supports batch export and ZIP packaging.

Universal data ingestion — loads CSV, Excel, JSON, and Parquet files into unified DataFrames with auto-detection of format, encoding, and date columns. Supports glob patterns for batch loading.

Transform and enrich tabular data — numeric normalization (min-max, z-score, robust scaling), financial metric calculation (revenue, margin), time feature extraction, moving averages, growth rates, and derived column computation.

Validate tabular data against JSON schemas and configurable business rules — column presence, type checks, nullable/unique constraints, range rules, referential integrity against lookup tables, and temporal consistency.

End-to-end ETL pipeline for retail sales data — ingests CSV transactions, cleans and validates data against product catalog, computes derived metrics (revenue, margin, moving averages), aggregates by multiple dimensions, and exports results in CSV/JSON formats with a summary report.

Compute statistical summaries (mean, median, p95, p99, min, max) for numeric columns in a CSV file and output a JSON report.

Parse structured JSON-lines log files, filter by time range and severity, output as CSV.

Analyse text files for sentiment polarity (positive/negative/neutral) using keyword-based scoring.

Composite pipeline skill that chains log-parser and csv-stats-reporter to filter JSONL logs and compute response-time statistics in a single step.

Parse CSV files into structured records with configurable delimiter and encoding.

Data loading stage — parses CSV and JSON files into normalized records with configurable delimiters and encoding detection.

Compute derived columns from existing data using arithmetic operations (sum, diff, ratio, multiply).

Filter rows from a JSON records file by column comparison expressions.

Data transformation stage — computes derived columns and filters rows by conditions.

Format and write final report from stats aggregation results in CSV or JSON format.

Reporting stage — aggregates statistics by group and writes formatted output reports.

Compute aggregate statistics (mean, median, sum, count, min, max) grouped by a column.

End-to-end CSV data analysis and report generation pipeline — loads multi-format data, computes derived columns, filters, aggregates statistics, and exports formatted reports.

Validate tabular data against configurable business rules — range checks, cross-column constraints, referential integrity against lookup tables, and temporal consistency.

Load and parse CSV files into structured data frames with configurable encoding, delimiter, and date parsing options.

Export pandas DataFrames or processed datasets to well-formatted CSV files with configurable encoding, delimiter, quoting, and column ordering.

Comprehensive data cleaning toolkit — handles missing values, duplicate removal, type coercion, whitespace trimming, and outlier detection in a single pass over the dataset.

Identify and remove duplicate records from tabular datasets using exact-match, fuzzy-match, or composite-key-based deduplication strategies.