Truero

Different jobs need different checks. Training an ML model adds machine-learning checks (data leakage, drift, scale). Cleaning or exploring runs the data-quality checks. Compliance runs both — plus privacy and ethics checks.

Some industries — healthcare, finance, insurance — are treated as high-risk under the EU AI Act. If you pick one of these, Truero applies stricter thresholds and flags regulatory risks.

Compliance data gets extra checks for personal information (PII) and regulatory risk, with tighter bias and missing-data thresholds. Answer "No / Not sure" if none of this applies.

Your goal decides which checks matter most. "Understand my data" and "Improve data quality" focus on data checks. "Prepare for model training" adds machine-learning checks. "Compliance report" adds the ethics and regulatory checks.

Different models have different data needs. Neural networks need lots of rows and well-scaled features; tree models are more forgiving. Truero tailors its size and scaling advice to your choice. Pick the closest match or "Skip" if unsure.

The "target" is the outcome your model should predict — for example "churned" or "fraud". Truero checks that no other column accidentally gives the answer away (data leakage).

If your data contains a column that groups people — gender, age band, ethnicity — Truero can check that every group is fairly represented (bias audit). Skip this if you don't have such a column.

A one-line description helps you recognise this scan later in your dashboard history — it is shown alongside the results.

Tags are labels you attach to this scan — like project name, environment (test/live) or team. They make it easy to find this scan later in the dashboard search.

Extra checks you choose yourself for specific columns — for example "email must look like a real email" or "no blank values in the phone column". Pick a column, pick a rule, add as many as you like. They run alongside the standard checks.