I’m Tomonori Kawai, a data scientist and data engineer at Japan’s largest free ad-supported streaming platform. My work spans the full stack, from the statistical models to the production systems that deliver the ads.
The core is statistical modeling for ad measurement: Brand Lift Study methodology, causal inference for ad effectiveness, and cross-campaign analysis over individual-level outcome data. Around that, I build forecasting models in Python, the analytics warehouse with dbt on BigQuery, the cloud infrastructure in Terraform, and the ad delivery services in Go.
After a master’s in molecular biology at Kobe University, I worked in fragrance R&D, healthcare, and insurance before moving into ad-tech. This blog covers the statistical methods behind advertising measurement, with working implementations.