Potthast, JonasGrimm, ValentinRubart, Jessica2023-08-242023-08-242023https://dl.gi.de/handle/20.500.12116/42144With the need to explain complex machine learning (ML) models it is necessary to explore human friendly visualizations and interaction techniques of data. In our position paper, we discuss a unique way for interacting with machine learning data to help decision makers in developing a mental model. This can increase trust towards the ML models and make the complexity digestible. In our approach, we combine visual analytics with explainable AI. In particular, we present a Mixed Reality based scatterplot application making use of SHAP values, where feature axes are automatically adjusted when diving deeper into parts of the data.Text/Workshop Paper10.18420/muc2023-mci-ws16-389