// Data Analyst · Statistician · Melbourne

Statistical depth
meets business impact.

I'm Yoonseul Choi, a data scientist with three years in healthcare analytics at HD Hyundai Medi Plus and an MS in Applied Statistics from Hanyang University. My work covers KPI design and Power BI dashboards on the industry side, and Bayesian hierarchical and spatial modelling on the research side.

Open to roles in Australia Partner Visa · No work-rights cap Healthcare · Public Sector · Spatial
// 01 · About

Statistician by training, analyst by practice.

Based
Melbourne, AU
Origin
Seoul, KR
Field
Spatial · Bayesian · Healthcare
Industry
3 yrs · Digital Healthcare
Languages
EN (Advanced) · KR (Native)
Status
Open to AU roles

My background is in statistics, with a graduate focus on spatial and Bayesian methods. After my BSc at Dongguk University, I joined HD Hyundai Medi Plus, the digital healthcare arm of HD Hyundai, where I spent three years on applied analytics across consumer, B2B, and clinical product lines.

The work split roughly into three areas. First, defining what an active member looks like for a digital health app and operationalising KPI metrics, retention, and churn through Power BI dashboards used in weekly executive review. Second, building an end-to-end healthcare cost prediction model on five years of Korea Health Panel data, using random effects regression layered with ensemble learning. Third, partnering with Kyobo Life Insurance on the clinical validation and B2B dashboarding for their Employee Assistance Program launch.

In parallel I'm finishing my MS in Applied Statistics at Hanyang University under Prof. Jungsoon Choi, with two arXiv preprints on spatial optimisation, a research internship at Academia Sinica on nonparametric bootstrapping for spatial point patterns, and a Workshop on Spatial Statistics 2025 talk.

I'm now looking for analyst, data scientist, or statistician roles in Australia. Open to most domains where rigorous quantitative work translates into real business or policy decisions.

// 02 · Selected Projects

Four projects, four different kinds of evidence.

P/01 HD Hyundai Medi Plus · 2024

KPI Metrics & Member Funnel

Defined what counts as an "active member," built conversion funnels, and quantified retention, churn period, and carrying capacity for a digital health app, feeding a Power BI dashboard used in weekly executive review.

90%Dashboard contribution
5+Member segments
WeeklyExec cadence
Power BI Clustering Funnel PostgreSQL Python
Read case study →
P/02 HD Hyundai Medi Plus · 2024

Healthcare Cost Prediction Model

Five-year panel study on Korea Health Panel diabetes cohort. Random Effects Linear Regression on log-cost, four derived specifications compared, final 19-variable model validated on multicollinearity and autocorrelation. Productised as a member-facing lifestyle savings calculator.

2014–18KHP panel
19 varsFinal model
−11.5%10-yr cost · 1 min/day exercise
Python statsmodels Random effects Panel data Log transform
Read case study →
P/03 Kyobo Life Insurance B2B · 2024

Employee Assistance Program: Clinical Validation

Insight derivation for Kyobo Life Insurance's EAP launch. Ran user analysis and chronic-disease management effectiveness across treatment and comparison cohorts, then built B2B dashboards that became the foundation for the commercial pitch.

B2BStakeholder
ClinicalEvidence
StrategyOutcome
Power BI Causal inference SQL Cohort analysis
Read case study →
Optimal mining site map showing Boeun-gun, Okcheon-gun, Yeongdong-gun with predicted optimal machine counts and annual net profits
P/04 arXiv 2511.21773 · WSS 2025

Spatial Two-Stage Hierarchical Optimisation for Bitcoin Mining Sites

A two-stage hierarchical spatial framework identifying where surplus solar electricity in South Korea makes Bitcoin mining nationally profitable. Cost-benefit optimisation at the regional level, then geographic constraints applied at finer resolution. Three optimal regions identified, with combined potential of around USD 307M in annual net profit. Presented at WSS 2025; available on arXiv.

3 regionsOptimal sites
~$307MCombined annual net profit
2 preprintsarXiv
Spatial stats Optimisation R Hierarchical model Sensitivity analysis
Read case study →
// 03 · Toolkit

The stack I reach for, most days.

Languages

  • Python daily
  • R daily
  • SQL · PostgreSQL daily
  • MySQL fluent
  • SAS · SPSS academic

Visualisation

  • Power BI 90% contrib.
  • Tableau fluent
  • Amazon Quicksight fluent
  • ggplot2 · matplotlib daily
  • Excel advanced

Statistics & ML

  • Bayesian hierarchical research
  • Spatial statistics research
  • Random / mixed effects applied
  • CatBoost · LightGBM applied
  • Tensorflow / Keras applied

Workflow

  • Git · Bitbucket daily
  • NLP · NLTK · KeyBERT applied
  • Stakeholder reporting 3 yrs
  • STAR-style writeups routine
// 04 · Experience

Roles to date.

Jul – Aug 2025

Research Intern

Institute of Statistical Science, Academia Sinica · Taipei

Under Prof. Junho Yang, working on nonparametric bootstrapping for spatial point patterns, focusing on inference for point processes with measurement error and computationally efficient methods for very large datasets. Concluded with a poster presentation.

Jun 2022 – Apr 2025

Data Scientist / Analyst

HD Hyundai Medi Plus Solution · Seoul

Healthcare analytics across consumer, B2B, and clinical product lines. Owned the KPI committee, built the cost prediction model end-to-end, and partnered on the Kyobo EAP launch.

  • Power BI dashboards across executive, B2B, and internal use cases (90% contribution on flagship)
  • PostgreSQL automation for healthcare lifelog ETL
  • Internal data dashboards on Amazon Quicksight
  • NLP keyword analysis (NLTK, KeyBERT) for product feedback
  • Weekly academic-paper review and internal technical posts
Oct 2023 – Jan 2024

Analyst, National Fitness 100

Korea Institute of Sports Science (KISS)

Investigated the representativeness of the National Fitness 100 dataset against the Korea National Health and Nutrition Examination Survey (KNHANES).

Nov 2022 – Jan 2023

Analyst, Handball Player Selection

Korea Institute of Sports Science (KISS)

Multiple linear regression and Random Forest to identify promising handball players at an early stage. Handled data imbalance, KMO test, and Cronbach's alpha. Random Forest classifier reached F1 = 0.9936.

// 05 · Research & Talks

Publications, talks, and posters.

Paper Nov 2025

Spatial Two-Stage Hierarchical Optimization Analysis for Site Selection of Bitcoin Mining in South Korea

Choi, Y. & Choi, J. (2025). arXiv:2511.21773.
arXiv ↗
Paper May 2025

Leveraging Surplus Electricity: Profitability of Bitcoin Mining as a National Strategy in South Korea

Choi, Y., Jeong, J., & Choi, J. (2025). arXiv:2505.00303.
arXiv ↗
Poster Aug 2025

Nonparametric Bootstrapping for a Spatial Point Pattern

Academia Sinica Summer Internship Poster Presentation, Institute of Statistical Science.
Poster ↗
Talk Jul 2025

Bitcoin Mining as a National Strategy: Spatial Optimisation

WSS 2025 Workshop on Spatial Statistics, Academia Sinica.
Workshop ↗
Mini Thesis Apr 2025

Lifestyle Patterns Comparison Between Individuals with Dementia and Without, Using Lifelog Data

Hanyang University, Department of Applied Statistics.
Paper ↗
Mini Thesis Aug 2024

Bayesian Hierarchical Multivariate Spatial Modelling on the Prevalence of Diabetes and Hypertension in Korea

Hanyang University, Department of Applied Statistics.
Paper ↗
// 06 · Education

Academic background.

MSc · 2024 – 2026

Applied Statistics

Hanyang University · Seoul, Korea
4.31 / 4.5GPA
Spatial · BayesianFocus

Under Prof. Jungsoon Choi (Spatial Statistics Lab). Coursework spans Bayesian Statistics, Spatial Statistics, Statistical Computing, Theory of Linear Models, Time Series, and Statistical Data Science.

BSc · 2017 – 2022

Statistics

Dongguk University · Seoul, Korea
4.13 / 4.5Overall GPA
4.30 / 4.5Major GPA

Graduated in seven terms. Multiple academic excellence scholarships, plus tutoring assistant for Mathematical Statistics I. Member of the Big Data Society academic society 2019–2020.

// 07 · Contact

Let's talk.

Currently looking for analyst, data scientist, and statistician roles across Australia. Open to most domains where rigorous quantitative work translates into real business or policy decisions.