Marketing data analyst, 12+ years in. Based in Madeira, Portugal — working remotely with teams across Europe.

I spend most of my time turning messy ad and product data into things people actually act on. Attribution models that tell you which channels work. Dashboards that surface the right metric at the right grain. Pipelines that make the analysis possible at all.

The signature project on my CV: unifying two completely separate post-acquisition data ecosystems into a single BigQuery warehouse, then building traffic-source-specific attribution on top of it. The dashboards from that work drove product decisions for over a year.

GA4 + BigQuery (where most of my SQL lives), server-side GTM, AppsFlyer, dbt, Looker Studio, Power BI. Meta Ads, Google Ads, Yandex Direct on the buying side. n8n for the glue work.

• ga4-attribution-models — eight attribution models in BigQuery SQL: last-click, first-click, linear, time decay, position-based, last non-direct, data-driven (BQ ML), and a cross-model comparison. Runs on Google's public GA4 sample.
• bigquery-meridian-mmm — marketing mix model using Google's Meridian library. Bayesian inference, full posterior distributions, geo-level support, MCMC diagnostics.
• simple-marketing-mix-model — the OLS + adstock + Hill saturation version, for when full Bayesian is overkill.
• marketing_analytics_sample_reporting — dbt project unifying Facebook, Google, and TikTok ad spend into a clean reporting mart.
• landing-page-ab-testing — SQL templates for AB analysis: frequentist t-test, Bayesian, SRM detection, sample-size calculator, guardrail checks.
• cohort-log-predict — predict day-30, day-90, day-365 retention from just two data points using a power-law fit.

I run portugalevents.eu — a small platform helping local events in Portugal get found. n8n + GA4 + BigQuery + Looker Studio.

Open to marketing data analytics roles, remote. LinkedIn.