Oleg Muratov
Data Science & Causal Inference
Economist (PhD, Penn State) applying causal inference, experimentation, and the SQL/Python stack to real-world data. This is my data-science portfolio — for my academic profile and CV, see olegmuratov.com.
Skills
- Causal inference
- Experimentation / A‑B testing
- Regression · panel fixed effects
- Python
- SQL · DuckDB
- pandas · scikit‑learn
- ETL & data pipelines
- Stata
- Git
Projects
Google Merchandise Store — GA4 Funnel, Channels and Week-over-Week Tests
Rebuilt 360,129 visits from 4.3 million raw events in Google's public GA4 export, and audited what the export can be trusted for: the store is its own traffic source on 53% of source-tagged events, some purchases are recorded up to four times, and add-to-cart is missing on 18 days. Conversion rose to 2.0% in early December and fell to 0.56% in early January; 7 of 12 week-over-week changes are beyond chance, while no swing in traffic stands out against the quarter's typical variation. Referral visits convert at 4.1%: 7.7% of visits, but 23.5% of visits with a purchase.
London Airbnb — How Posted Prices Behave Over the Calendar
How do 85,000 London Airbnb listings price the year ahead? About a third post one price on every date. Within the same listing, posted prices are 2.4% higher on Friday and Saturday nights and about 11% higher over Christmas and New Year, and professional hosts' listings show roughly twice the premium. Built on posted prices only, with no booking or revenue claims the data can't support.
Certifications
