OpenCLD: A Scalable Python Library for System Dynamics Modelling

The practical application of System Dynamics (SD) modeling is constrained by proprietary software that is expensive, inflexible, and poorly integrated with modern data science ecosystems. Existing open-source alternatives often lack the features and native development capabilities required for contemporary analytical workflows. This paper introduces OpenCLD, a lightweight, open-source Python library designed to bridge this gap by providing a native and intuitive framework for SD modeling. Built on object-oriented principles, OpenCLD enables the construction of complex models using standard SD components while ensuring seamless compatibility with the scientific Python ecosystem, including pandas, scikit-learn, and TensorFlow. The library supports modern development practices such as interactive modeling in Jupyter notebooks, containerized deployment, and integration with machine learning pipelines. OpenCLD democratizes system dynamics, offering a powerful and accessible alternative for developing transparent, reproducible, and scalable models.