Professional Portfolio
Sangmin Lee
Email · aquariusmin01@naver.com
GitHub · https://github.com/aquariusmin
Portfolio page · https://www.abovemin.com/en/portfolio
I test the assumptions behind data and turn it into metrics a decision can rest on.
The featured cases are the Busan urban-rail dwell analysis, telecom churn XAI, and satellite night-light GDP analysis. Exploratory work such as the Arctic route study is placed lower as supporting material.
Markets
International trade, macro context, alternative data, and country or industry screening.
Finance
Financial statements, valuation logic, sensitivity checks, and assumption audits.
Analytics
Python, statistics, regression, classification, explainable AI, and survey analysis.
Execution
Dashboards, reports, MVP flows, operating checks, and clear next-step recommendations.
Case study summary
Three featured projects and Explore work
Case 01
Public Data · Urban Policy
Busan Station Dwell Conversion Analysis
Question · What does a ridership-based policy target list miss? Are alighting and dwelling the same thing?
Decision value · Ranking stations by failed dwell conversion instead of raw ridership produces a different target list for commercial-district renewal budgets, and producing that list is the point of the analysis.
Boundary · Dwelling population and dwell time are estimated from gate records; they are not measured mobility data from telecom or card sources.
Case 02
Customer Analytics · Strategy
Telecom Customer Churn Analysis
Question · Which customers are most likely to leave, why are they at risk, and which retention response fits each segment?
Decision value · The analysis turns a binary prediction task into a prioritization framework and proposes retention strategies targeting a 5.0 percentage-point reduction in churn; the target has not been achieved or validated.
Boundary · The 5.0 percentage-point churn reduction is a proposed planning target, not an achieved business result.
Case 03
Economics · Alternative Data
Satellite Night-Light GDP Analysis
Question · When official statistics are scarce or late, how far can night-light intensity serve as a supporting indicator of economic activity?
Decision value · Rather than replacing official statistics, night-light data can serve as a supporting signal for first-pass screening of countries and regions, or for monitoring change where statistics are missing.
Boundary · R-squared = 0.819 describes explanatory fit in the simple model; it is not 81.9% prediction accuracy.
Case A01
Climate Data · Logistics Strategy
Arctic Route Accessibility Analysis
Question · How much has sea-ice loss changed the operational window of Arctic routes, and what choice does that leave a Korean shipper?
Decision value · Reduced the finding to two decision-grade numbers, navigable months per year and nautical miles saved, so routes can be compared on when and for how long they are usable.
Boundary · Navigable months describe physical accessibility under concentration thresholds; they do not mean commercial transit is operationally viable.
Case A02
Fintech · Strategy Validation and Operations
Quant Trading Automation
Question · What does running trading rules as a service require beyond strategy code, and how do you confirm the rules actually have an edge?
Decision value · Instead of claiming returns, set out in numbers and operating controls the conditions under which a strategy may run and the signals that should switch it off.
Boundary · Runs on paper and mock accounts and on small real-money accounts, labelled bot by bot on /lab. Backtest figures are simulations on historical data, not a live record, and claim no return, win rate, or capital growth.
Case A03
Financial Research · Strategy
Korean Air Financial Analysis
Question · How should Korean Air be evaluated when post-pandemic recovery, large investment needs, leverage, and uncertain cash-flow assumptions cause valuation methods to diverge?
Decision value · When valuation methods diverge, decision-makers need an assumption audit rather than a mechanical average or an attractive headline number.
Boundary · This is conditional classroom analysis, not investment advice or a verified public target price.
Case A04
Applied Analytics · Model Selection
Financial AI Model Study
Question · How do preprocessing, regularization, model complexity, and hyperparameters affect validation and test performance?
Decision value · Use additional complexity only when validation evidence justifies it; a more advanced method is not automatically the stronger business choice.
Case A05
Service Planning · MVP
Phorage Brand and Commerce MVP
Question · How can a personal photography concept be translated into a testable product and service experience before public launch?
Decision value · The project demonstrates service planning, MVP prioritization, physical-digital integration, and the definition of launch hypotheses and metrics.
Boundary · Pre-launch MVP only. No public launch, public traction, large-scale orders, or verified sales metrics are claimed.
Case A06
Research Design · Statistics
Blood Type and Personality Survey Study
Question · Does a familiar social belief about blood type and personality remain supported when translated into hypotheses and tested with survey data?
Decision value · The study demonstrates how to turn a popular claim into testable hypotheses and restrict the conclusion to what the evidence supports.
Boundary · One isolated significant result does not prove a broad relationship between blood type and personality.