Data foundations, research records, and tools for reviewable decisions.
Research systems · All repositories
I build auditable data and research systems for China equities, index options, commodities, and cross-asset portfolio review. Three data engines connect source snapshots and quality checks to research reports, portfolio evidence, and capital analysis. Source dates, transformations, caveats, and results stay connected. Alongside this work, I explore reproducible research and interactive creative tools.
|
Data foundations A-share, CFFEX derivatives, and commodity datasets with point-in-time boundaries, provenance, coverage checks, and explicit freshness states. |
Market research Cross-asset report archives, regime views, and options research that separate source evidence from interpretation and preserve each report’s status. |
Portfolio & capital Portfolio-review workflows, solvency scenarios, and IFRS 17 learning tools that connect management questions to traceable calculations and evidence. |
|
Global Cross-Asset Radar Grok morning/evening research archives across macro, equities, rates, commodities, FX, and options. China commodity reports and original records remain in the original radar archive; report dates and validation states remain explicit. Python · Reports · Cross-asset research |
China Stock Engine An A-share daily-data foundation and price-volume discovery engine with reproducible BREAKOUT/WATCH screens, source-date checks, and coverage controls. Python · PIT data · Technical screening |
|
China Options Engine An engine that turns CFFEX futures and option-chain data into implied volatility, Greeks, gamma exposure, and daily state. Python · CFFEX · IV and Greeks |
China Commodities Engine China commodity futures EOD, separate night-session snapshots, physical-market inputs, and sampled options data with explicit provenance, coverage, and readiness states. Python · Commodity futures · EOD |
|
Portfolio Review A monthly portfolio-review workflow from Excel snapshots to traceable Parquet releases, analytics, and asset-level evidence. Python · Parquet · Portfolio analytics |
Solvency An insurance solvency and asset-allocation workbench for historical review, scenario shocks, target solving, and capital attribution. Python · Streamlit · Scenario analysis |
- Ma Capital OS — a read-only multi-asset research platform in development. Empirical research acceptance remains incomplete, and recommendations are blocked.
- IFRS 17 Learning Workbench — an offline teaching model for contract roll-forwards, VFA, IFRS 9/17, and financial-statement reconciliation, using synthetic cases.
- Options OI Monitor — short-window unusual open-interest monitoring with explicit coverage limits; OI alone does not establish trade direction.
Liber Primus Lab explores Cicada 3301 through reproducible rune-level experiments, fixed sources, and recorded failures. Guanbian / 观变 is an I Ching-inspired reflection and review prototype. PulseForge Mania turns local audio analysis and reproducible pattern rules into editable osu!mania charts.
The website catalog connects the wider project collection. The Google Drive archive is paused and retained for historical reference.
Evidence before signal. Keep source dates, transformations, assumptions, and data gaps visible. Build the smallest useful system that can still be challenged and rechecked; keep sensitive data local where appropriate.
Research and educational software. No repository here is an order-generation system or investment advice.