Minute-level meteorological dataset for lake-breeze nowcasting along eastern Lake Michigan, July 2024

This dataset contains minute-level meteorological observations and processed modeling inputs used for lake-breeze nowcasting along the eastern shore of Lake Michigan during 1–31 July 2024. The dataset supports the development, evaluation, and interpretation of machine-learning models for lake-breeze prediction, including XGBoost, LightGBM, Logistic Regression, persistence baseline comparisons, SHAP interpretation, and feature ablation experiments. The uploaded archive includes the raw merged dataset, the autocorrelation-based quality-controlled dataset, the quality-controlled and lake-breeze-labeled dataset with the date column retained, and the final date-free model input dataset used for hyperparameter optimization, model training, SHAP analysis, and ablation experiments. The binary lakeBreeze label uses 1 to denote lake-breeze occurrence and 0 to denote non-lake-breeze conditions.