Model Presentations
Selected model presentations, dashboard links, and workflow diagrams from publications and the K-WaterGuard AI Agent, designed for collaborators, students, research groups, and climate-tech partners.
Research Models
Hybrid deep learning framework for precipitation forecasting with integrated large-scale climate drivers.
Machine learning workflow for downscaling and projecting rainfall patterns using CMIP6 climate model data.
Graph-based climate-crop modeling framework for analyzing climate variability impacts on crop yields.
Direct and indirect potential evapotranspiration prediction for river-basin water planning.
Machine learning and deep learning model families for hydrology, streamflow, and basin-scale environmental prediction.
Hybrid deep learning and time-series trend analysis for detecting climate-change signals in precipitation and temperature records.
Interactive Options
Open the deployed dashboard to inspect water-quality monitoring outputs, maps, alerts, and daily evidence generated from public API data.
Launch dashboardFor proprietary or project-specific models, collaborators can request a scoped Streamlit or Gradio prototype using their station, CMIP6, GIS, or farm datasets.
Request prototype demoReview the model cards above for data ingestion, feature engineering, validation, and deployment logic before deciding on a technical consultation.
Discuss architectureModel Workflow
Weather stations, remote sensing, CMIP6, NWP outputs, basin data, crop records, and water-quality observations.
Climate drivers, lag variables, spatial attributes, trend indicators, and quality-controlled predictors.
LSTM, RNN variants, WaveNet, graph neural networks, machine learning, and hybrid frameworks.
Performance metrics, benchmark comparisons, error analysis, explainability, and scientific interpretation.
Forecasts, GIS risk maps, publication figures, dashboards, reports, and climate-smart recommendations.
Agentic System
K-WaterGuard AI is a Python-based agentic water-quality monitoring system for South Korea. It collects public water-quality API data, stores historical records, creates plots and maps, generates an HTML dashboard, and supports optional chatbot integration through a secure backend.