Tacke Genannt Unterberg, LeahKoren, Istvánvan der Aalst, Wil M.P.Michael, JudithWeske, Mathias2024-02-192024-02-192024978-3-88579-742-5https://dl.gi.de/handle/20.500.12116/43613Data interoperability in Industry 4.0 is a continuous challenge for industry and research. Many organizations face the challenge of managing data lakes that, without proper governance, risk becoming disorganized `data swamps' with disparate data models and formats. This heterogeneity leads to inefficient data utilization.Standardization efforts have produced suites of extensive models as they try to accommodate diverse requirements while still being comprehensive. Their complexity has hindered their adoption. To address this, we propose a minimal intermediate meta model for a frequently considered type of data in smart manufacturing, namely Machine Data. This type of data is central to industrial IoT platforms and research efforts on Digital Shadows & Twins. It encompasses raw time series and event data from sensors and digital controllers. This model-in-the-middle is intended to bridge the gap between heterogeneous source systems and highly structured and semantically clean input for data science techniques. To be broadly applicable, it has to be minimal and favor abstraction over details. We equip it with a standardized exchange format based on CSV, which reduces friction in data sharing. Furthermore, we provide a precise mathematical formalization that connects it to the language of data science methods. This enables the generic implementation of methods that can easily be reused and combined. Finally, we validate the model together with initial tool support in the large-scale cluster of excellence Internet of Production (IoP). We conclude that it is possible and feasible to accelerate the realization of the ambitions for the future of manufacturing using such minimal models.enIndustry 4.0machine dataexchange formatText/Conference Paper10.18420/modellierung2024_0111617-5468