Data Scientist · Machine Learning · Alternative Data
I build statistical and machine learning models on alternative data, along with the end-to-end data and modeling pipelines that take them to production. Most recently in real estate investing; before that, utilities and cancer research.
Shaped by fast-paced startup work: independent, self-driven and quick to learn a new domain.
Case studies from real estate investing, utilities and data for good. Charts are illustrative and use synthetic data.
One system, built end to end: ranks every US zip code by expected return and simulates the portfolio.
A beta-regression market ranking plus a survival model for how fast vacant buildings lease.
A rare-event model ranking 811 tickets so field teams inspect the riskiest digs first.
Privacy-safe small-area units for Denmark, built from road networks.
More Paper · BMC Medical Research Methodology (2020) ↗ Talk · Spatial Data Science Conference (2022) ↗ Earlier school projects →
Also built Twin-market finder (dynamic time warping) · Shared geospatial and Kalman-smoothing utilities · REIT same-store NOI signal testing · LLM research harness
Eight years building forecasting and geospatial models, most recently for real estate investing.
Vice President, Data Scientist (2024–present); Data Scientist (2021–2023). Industrial and single-family investment models, alternative-data research, team tooling.
Senior Data Scientist; Data Scientist. Damage-risk classification and time-series forecasting for utilities.
Geospatial Data Analyst. Spatial and survival models on social determinants of cancer outcomes.
Master of Urban Spatial Analytics (2018); B.S. Landscape Architecture (2017).
© 2026 Yinuo Yin