001
Platform Adoption Model
ML and causal inference models that drove 5–10% regional adoption rate increases for an online B2B platform.
Machine learning, causal inference, and AI/LLMs — turning messy real-world data into systems that drive measurable impact.
View WorkI'm Hannah Lyon, an AI-native data scientist based in Florida, USA. I specialize in machine learning, causal inference, and AI/LLMs — translating messy, real-world data into systems that drive impact.
My background blends statistics and software engineering. Before diving into data science full-time, I studied economics at NYU in Shanghai, where I developed a deep appreciation for data and econometric research.
When I'm not wrangling datasets, you'll find me streaming or napping with my dog, Molly. I believe the best data science is done with as much empathy as it is with code.
ML and causal inference models that drove 5–10% regional adoption rate increases for an online B2B platform.
Improved dataset quality by 20% by contributing to data modeling, MDM, and ETL pipelines powering downstream feature engineering and segmentation.
Facebook Prophet model forecasting monthly store sales for the next half-year, surfaced in an interactive portfolio dashboard.
NLP-powered fuzzy matching tool using TF-IDF that eliminated hours of manual data labeling for recurring ingestion workflows.
Graph-based collaborative filtering model that spearheaded two personalization initiatives on QVC's e-commerce platform, boosting conversion by 2%.
Data-driven attribution model that saved 15% of marketing spend at Cuyana, paired with a SQL/Tableau dashboard tracking daily payment method behavior.
Have a project, collaboration, or just want to talk data? I'm always open to interesting conversations.