An end-to-end predictive analytics pipeline built in R using 30-minute interval data from AEMO. This project utilizes a Fourier-expanded ARIMA model to capture extreme intraday solar generation volatility. By modeling these daily cycles, the pipeline drops the baseline forecasting error (MAPE) from 35.8% down to 4.73%. Includes an embedded interactive Tableau dashboard tracking the 48-hour forward projections.

An interactive geospatial analysis evaluating the intersection of extreme thermal stress (Urban Heat Island effects) and socioeconomic vulnerability across metropolitan Adelaide, South Australia. Utilizing R, Quarto, sf, and leaflet, this project moves beyond standard standalone heat mapping to execute a 3x3 bivariate matrix and an ABS suburb boundary spatial join, pinpointing exact localized neighborhoods requiring urgent climate-resilience and infrastructure intervention.