BIM-Based Dataset for Machine Learning Energy Prediction

This dataset contains Building Information Modeling (BIM)-based simulation data generated by integrating Autodesk Revit analytical models with Autodesk Insight energy simulations for office buildings located in Orlando, Florida (ASHRAE Climate Zone 2A). The data were used to train and validate Machine-Learning (ML) surrogate models for predicting building energy performance, specifically Energy Use Intensity (EUI) and Operational Energy (OE). The dataset includes parametric variations of geometric, envelope, system, and operational parameters derived from a reduced-factorial Design of Experiments (DOE). It is provided in both Excel (.xlsx) and Attribute-Relation File Format (.arff) structures for compatibility with Weka and other ML platforms. This dataset supports the publication: “BIM-Based Machine Learning Framework for Early-Stage Building Energy Performance Prediction” (submitted to Applied Sciences, 2025).