{"repo":"antonbarr-data/dqt","free":true,"listed":false,"github":"https://github.com/antonbarr-data/dqt","clone":"git clone https://github.com/antonbarr-data/dqt.git","description":"A Data Quality Tool for Agentic BI that tells you the what and the why","language":"Python","stars":12,"topics":["ai-bi","business-insights","business-intelligence","causal-discovery","causal-inference","data-quality","intelligent-agent","semantic-layer","agentic-bi"],"license":"MIT","category":"analytics","readme_excerpt":"質 dqt Data Quality Tool for Agentic BI. The what and the why. 64 detectors across 5 families (drift, outlier, time series, distribution, rule) · best F1 0.933 (holt winters / wasserstein 1) · full results dqt is a data quality tool built for Agentic BI : the world where agents, not just people, read your metrics and write your reports. It connects to your warehouse, imports your semantic layer from open formats (Google OKF and Apache Ossie), and wraps everything in a statistical quality harness with column-level lineage and causal explanations. The result is a grounded semantic layer you did not have to hand-write, plus guardrails that keep the metrics and reports your agents produce correct, explainable, and on-spec. Why dqt is different: every other data quality tool tells you that a metric broke. dqt is an open, gap-free superset that also tells you why - 30+ statistical detectors, column-level lineage, and a causal layer (Granger, PCMCI+, confirmed DAGs with Shapley attribution) that Great Expectations, Soda, Elementary, Dataplex, and Monte Carlo do not offer. Full comparison. --- The problem it solves Without dqt: orders.amount null fraction = 0.05 -- threshold exceeded . Now what? Go dig through git log, dbt docs, warehouse history. With dqt: --- Four layers - Statistical detectors - MAD, double-MAD, isolation forest, KS, STL residuals, adjusted boxplot fences. Plus completeness, validity, freshness, schema-change, and SQL-assertion checks. Every detector returns (verdi","default_branch":null,"files":null,"tree":[],"storefront":"/r/antonbarr-data","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/antonbarr-data/dqt/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}