Source-Resolved Machine Learning and Causal Inference Dataset for Summertime Ozone Formation in Metro Detroit

This dataset supports the analysis of source-resolved drivers of summertime ozone (O3) formation in Metro Detroit using observations from the 2021 Michigan-Ontario Ozone Source Experiment (MOOSE). It includes the machine-learning input dataset (MLinput.csv) containing ozone, meteorological variables, trace gases, spatial variables, and PMF-resolved VOC source contributions, along with a Jupyter notebook (Causal code.ipynb) implementing Double Machine Learning (DML) and Causal Forest (CF) analyses. The code estimates average treatment effects (ATEs) and conditional average treatment effects (CATEs) of individual VOC source factors on O3 after adjustment for measured confounding variables.