Final-year Data Science student at Deakin University. Most of my work starts with a question I'm curious about — then I build toward an answer through reproducible analytics pipelines, predictive models you can evaluate properly, and dashboards that make insights usable for non-technical stakeholders. Outside coursework, I mentor first-year mathematics students at Deakin and help organise campus events with the student association.
Technically, I'm comfortable across the full flow: ingestion, transformation, modelling, evaluation, and deployment. I work mainly in Python and SQL, with experience in machine learning libraries, BI and dashboard tools, relational databases, and cloud and DevOps tooling including Docker and GitHub Actions.
Open to data analyst, analytics engineer, and junior data scientist roles in Australia 🇦🇺 and Vietnam 🇻🇳.
Languages
Data & ML
Visualization & BI
Cloud, DevOps & Data Platforms
| Domain | Details |
|---|---|
| Data Analysis & SQL | Exploratory analysis, KPI reporting, PostgreSQL, DuckDB, SQLite |
| BI & Dashboards | Tableau, Power BI, Streamlit, Plotly for stakeholder decision support |
| Data Pipelines | Python + SQL ingestion and transformation; dbt-style modelling; GitHub Actions |
| Applied ML | Scikit-learn pipelines for classification and predictive modelling; model evaluation basics |
📊 NYC 311 Operational Analytics Pipeline
End-to-end operational analytics pipeline transforming raw NYC 311 service request data into actionable insights via a DuckDB-based transformation layer and a React + Vite dashboard, deployed with GitHub Actions CI/CD.
| Aspect | Detail |
|---|---|
| Stack | Python, DuckDB, React, Vite, GitHub Actions |
| Focus | Operational analytics & service-level reporting |
| Pipeline | Automated ETL → analytical tables → interactive dashboard |
| Deployment | CI/CD-driven, fully automated dashboard publishing |
| Repository | github.com/tuahung248 |
🩺 Fall Detection Classifier
A supervised ML classifier for detecting falls from sensor data, achieving 93% accuracy using a Random Forest model with careful feature engineering and evaluation.
| Aspect | Detail |
|---|---|
| Stack | Python, Scikit-learn, Pandas |
| Model | Random Forest (93% accuracy) |
| Focus | Health/safety monitoring use case |
| Repository | github.com/tuahung248 |
