Data scientist and economist based in San Francisco Bay Area. I build tools that help people think more clearly and make better decisions. One of my recent project is voice-first interfaces for capture, reflection, and time tracking. Before this, I spent several years applying machine learning to food-security forecasting in development settings, with work published in Applied Economic Perspectives and Policy and World Development.

• brief-voice-memo — voice-first time tracker. Whisper transcribes recordings; GPT-4o-mini auto-categorizes them (TODO / IDEA / QUESTION / REMINDER) and runs AI-powered daily and weekly coaching audits.
• dialect_voice_memo_wechatApp — the same idea, packaged as a WeChat mini-program for voice-driven capture and content organization.
• church-map - a location based church discovery app for searching churches by city/state, viewing them on a map, and reading or leaving reviews

• churn_prediction_music_box — calibrated churn-risk scoring tied to a unit-economics decision curve. Tells marketing what success rate clears break-even and what audience maximizes net retention value. Four models behind a unified inference layer.

• FoodSecurityPrediction — replication code for Machine learning for food security: Principles for transparency and usability (Applied Economic Perspectives and Policy).
• WD-Early-Warning-Food-Insecurity — replication code for A Data-Driven Approach Improves Food Insecurity Crisis Prediction (World Development).

• Email — zhou@yujun.net
• Portfolio:yujun.net
• Product: time.yujun.net