Олена

Olena Polishchuk
Data Analyst | Python, SQL, Power BI, Experiment Analysis
Sarzana, Italy | [відкрити контакти](див. вище в блоці «контактна інформація») | [відкрити контакти](див. вище в блоці «контактна інформація») | GitHub: github.com/olenapolishchuk

PROFILE
Data analyst with a strong quantitative background in Applied Mathematics and Mathematical Engineering. Hands-on portfolio
experience in A/B testing, statistical inference, SQL/BI reporting, dashboarding, data cleaning, and machine learning. Comfortable
translating analysis into practical product or business decisions, including when not to ship a change.

TECHNICAL SKILLS
Languages Python, SQL, MATLAB, C++

Analytics A/B testing, hypothesis testing, confidence intervals, Bayesian A/B add-on, cohort and funnel analysis

Python stack pandas, NumPy, scipy, statsmodels, scikit-learn, matplotlib, seaborn, openpyxl

BI and tools Power BI, Excel, Jupyter Notebook, Git, Google Sheets

ML methods Logistic Regression, Random Forest, SVM, Gradient Boosting, Neural Networks, model evaluation

Structured problem-solving, careful data validation, independent learning, clear communication of
Working style
analytical findings

SELECTED PROJECTS
Landing Page A/B Test Decision Analysis | Python, pandas, scipy, statsmodels, Bayesian inference
• Analyzed a 294k-user A/B test to evaluate whether a new landing page improved conversion.
• Cleaned and audited experiment data, checking assignment consistency, duplicate users, traffic balance and daily stability.
• Found no meaningful conversion uplift from the new page and no reliable segment where it performed better.
• Converted the analysis into a clear business decision: do not roll out globally due to negative expected impact.
Customer Churn Prediction | Python, scikit-learn, Telco dataset
• Built a churn prediction workflow with data cleaning, categorical encoding, scaling, stratified train/test split, Logistic Regression and
Random Forest models.
• Compared performance using classification reports, confusion matrices and ROC/AUC; Logistic Regression achieved AUC 0.836 and
outperformed Random Forest in this setup.
• Created feature-importance analysis to connect model output with likely churn drivers for business interpretation.
Sales Analytics and BI Reporting Portfolio | SQL, Power BI, Excel, Python |
• Created SQL reporting assets with schema scripts, insert scripts, PostgreSQL analytical queries and a Power BI report for revenue, AOV,
trends, top customers and products.
• Built sales and operations KPI dashboards covering revenue by region/category, monthly trends, on-time performance, average
completion time and operational status breakdowns.
• Automated HR reporting in Python/Excel, generating KPI outputs for employee status, salary, tenure and department-level turnover.
Academic Machine Learning Projects | Python, MATLAB, scikit-learn
• Phishing website classification: implemented and compared SVM, Random Forest, Gradient Boosting, Neural Networks and ensemble
methods on UCI structured web-feature data.
• Road traffic classification: prepared time-windowed signal features in MATLAB and trained SVM/RF models for traffic-state classification,
with reported accuracy around 98%.
• Used cross-validation, feature selection/PCA variants, confusion matrices and model comparison to evaluate reliability rather than only
reporting a single score.

EDUCATION
• Master's in Mathematical Engineering, University of L'Aquila, 2022-2025
• Master's in Applied Mathematics, Taras Shevchenko National University of Kyiv, 2022-2024
• Bachelor's in Applied Mathematics, Taras Shevchenko National University of Kyiv, 2018-2022

CERTIFICATIONS AND LANGUAGES
• SQL (Advanced) , HackerRank, 2025
• Languages: Ukrainian native; English intermediate (B1-B2)