This page is an overview of my data-science work, organised by deliverable and depth:
Working apps/pipelines with reproducible architecture and documented outputs.
Published:
A deterministic job-market intelligence system that turns messy postings into interpretable skill demand, salary signals, and clear best_now vs stretch recommendations — delivered as a reproducible Python pipeline + Streamlit app.
Published:
End-to-end SQL analytics project using the Lahman Baseball Database. Designed a complete relational workflow with schema creation, reusable views, advanced CTEs, window functions, and business-focused analyses on talent pipelines, salary dynamics, and player careers.
Peer-reviewed modelling/analytics work presented in a DS case-study format.
Published:
Developed a hierarchical Bayesian workflow to quantify partial altitudinal migration and system-wide community reshuffling across elevation and season in the Australian Wet Tropics.
Published:
Developed the analytical framework using generalised linear models to assess abundance changes of tropical species across forest gap gradients.
Published:
Developed and implemented a holistic Bayesian framework integrating microclimate, mechanistic physiology, biogeochemical processes, and population dynamics to identify causal pathways from climate change to survival and recruitment.
Published:
Revealed ecosystem cascades and biogeochemical pathways in tropical systems using Bayesian hierarchical modelling to quantify direct and indirect effects in complex ecological networks.
Published:
Applied hierarchical Bayesian models with satellite-derived predictors to identify climate-driven population changes in rainforest birds across space and time.
Published:
Developed Bayesian hierarchical models incorporating detection probability to forecast population viability and support elevated conservation status for imperilled species.
Published:
Developed a high-throughput spatial forecasting workflow of community turnover under climate change, optimising computational performance for multi-species forecasting across elevational gradients.
Published:
Used time-series GLMs and interactive visualisation (Shiny app) to nominate 14 bird species for elevated protection under national and international priority lists.
Published:
Applied ensemble machine learning to convert presence-only niche models into abundance predictions, combining big-data wrangling, advanced analytics, and multi-algorithm ensembles.
Short technique-focused builds used to cement specific skills.
Published:
Implemented core machine learning algorithms — from regression and classification to clustering and deep learning — through applied projects in Python. Focused on building intuition for model training, evaluation, and interpretability using Scikit-learn and Jupyter Notebooks.
Published:
Developed an applied EDA framework combining real-world case studies — emergency call records and financial time series — to demonstrate data wrangling, feature extraction, and visualisation workflows using pandas, seaborn, and plotly.
Published:
Developed a suite of applied Python mini-systems demonstrating the progression from procedural programming to object-oriented design and class interaction across real-world examples.