Dataset profiler & debugger v0.1.0b1
Lint your datasets before they reach training.
Catch critical dataset issues before they derail your ML pipeline. Profile, debug, and fix — all in one command.
$ pip install tablint
Successfully installed tablint-0.1.0b1
$ tablint scan train.csv --target Survived
Tablint Dataset Analysis
────────────────────────────────────
3 critical issues found
5 warnings detected
────────────────────────────────────
✗ high_missing_values Cabin: 77.1% missing
✗ class_imbalance Survived: ratio 1.58
✗ outliers Fare: 116 outliers
⚠ high_cardinality Ticket: 681 unique
⚠ feature_correlation SibSp ↔ Parch: 0.41Works with your ML stack
Profile, debug, and fix your datasets in seconds
Tablint catches critical issues before they derail your ML pipeline, explains the problems, and suggests context-aware fixes.
Intelligent Profiling
Detect missing values, skewed distributions, outliers, and data type inconsistencies across your entire dataset.
ML-Specific Checks
Identify data leakage, dataset drift, class imbalance, and high-cardinality features before training.
Automated Preparation
Get context-aware suggestions for encoding, imputation, scaling, and transformations — with auto-apply options.
Rich Reporting
Generate statistical summaries and exportable reports in HTML, PDF, Markdown, or JSON with embedded visualizations.
Production Pipelines
Output reproducible cleaning code and sklearn pipelines that integrate seamlessly into ML workflows.
CLI & Library
Use as a command-line tool for quick scans or as a Python library for deep integration into your codebase.
Three commands. Zero guesswork.
No need to change your workflow — bring your own tools, frameworks, and datasets.
Scan
Run a quick scan to get a summary of critical issues in your terminal. Add --target to enable ML-specific checks.
tablint scan dataset.csvAnalyze
Get comprehensive details about all detected issues including outliers, missing values, class imbalance, and data leakage.
tablint details dataset.csv --target SurvivedReport & Fix
Generate rich HTML/PDF reports with visualizations and auto-generated Python fix scripts and sklearn pipelines.
tablint report dataset.csv --format html --with-code03 / Usage
Use it your way
Quick terminal scan or deep Python integration — Tablint fits into any ML workflow.
# Quick scan for critical issues
$ tablint scan train.csv --target Survived
# Detailed analysis of all issues
$ tablint details train.csv --target Survived
# Generate HTML report with fix scripts
$ tablint report train.csv \
--format html \
--theme minimal \
--with-code
# Compare datasets for drift
$ tablint report train.csv \
--comparison test.csv \
--format html13 built-in checks. Every issue caught.
From basic data quality to ML-specific concerns, Tablint runs a comprehensive audit on your dataset.
Beautiful reports. Your choice of style.
Generate professional HTML reports with two built-in themes. Every report includes interactive visualizations and detailed diagnostics.
Minimal
Clean, professional look with soft colors and restrained typography. Ideal for stakeholder reports.
Neubrutalism
Bold borders, vivid accents, and high contrast. Stands out in presentations and dashboards.
Get started
Stop debugging datasets manually.
Install Tablint and get your first dataset audit in under a minute.
pip install tablint