Interactive machine learning (IML) enables models that incorporate human expertise because the human collaborates in the building of the learned model. Moreover, the expert driving the learning (human-in-the-loop-learning) can steer the learning objective, not only for accuracy, but perhaps for discrimination or characterisation rules, where isolating one class is the primary objective. Moreover, the interaction enables humans to explore and gain insights into the dataset, and to validate the learned models. This requires transparency and interpretable classifiers. The importance and fundamental relevance of understandable classification has recently been emphasised across numerous applications under the banner of explainable artificial intelligence. We use parallel coordinates to design an IML system that visualises decision trees with interpretable splits beyond plain parallel axis splits. Moreover, we show that discrimination and characterisation rules are also well communicated using parallel coordinates. We confirm the merits of our approach by reporting results from a large usability study.