About Dataxiom

Dataxiom turns datasets into reproducible, interactive explanations — connecting questions, evidence, models, and understanding.

What a study contains

Every published study states its question, documents where its data came from and under what licence, compares candidate methods against a baseline, reports how it was evaluated, and ends with what the result does not support. The last of those is not optional: a study cannot be published without a limitations section and at least one reference.

On model output

Where a study exposes a model, the model's version is shown alongside its output, and a score is presented as a probability only when the model has been calibrated. Everything else is labelled as an uncalibrated score, because that is what it is.

Relationship to Axiomizer

Dataxiom is Axiomizer's applied data-science layer: where Axiomizer explains a method, Dataxiom tests one.