Environmental Data Labeling and AI Modeling Lead
ClassificationI led the integration, preprocessing, and labeling of long-term air quality, meteorological, and chemical composition monitoring datasets for environmental AI modeling. I built data pipelines and used machine learning methods to classify pollution event types from spatiotemporal environmental data. I performed feature engineering and key factor identification to support source analysis and reduction strategies. • Applied Python and R for data integration, labeling of event types (dust, secondary pollution, firework emissions) using PMF, PCA, and ML models. • Established and iteratively improved annotation processes for pollution events and source regions in time-series and geospatial data. • Developed and validated predictive models, optimizing their accuracy via iterative relabeling and feature selection. • Leveraged SHAP and statistical approaches to label critical drivers and quantify their impact on pollution dynamics.