An Vo | Quantitative Economics & Data Science Portfolio

Data Analyst · BI Analyst · Junior Data Scientist

Quantitative Economics Master's Candidate @ University of Bonn

My Analytics Philosophy: I bridge the gap between rigorous econometric modeling and production-ready data pipelines. I focus on clean data governance, high-velocity analytical infrastructure (O(log n) efficiency), and deploying transparent, decision-ready statistical systems that empower executive corporate strategy.

🚀 Flagship Case Studies

🛒 1. FMCG Market Basket Studio

Transitioning traditional association rules into automated margin expansion.

• Objective: Transition traditional market basket association rules into an automated margin expansion engine for E-commerce recommendations.
• Methodology: Engineered an out-of-core FP-Growth pipeline (via DuckDB) to process high-volume transactions. Replaced standard statistical 'Lift' with a custom Expected Commercial Value (ECV) metric to prioritize gross margin impact over raw purchase probability.
• Delivery: Deployed an interactive Streamlit dashboard featuring NetworkX affinity graphs for category managers, alongside a Pydantic-validated FastAPI endpoint for real-time checkout integration. Includes a comprehensive Executive Summary PDF

⚖️ 2. Merger Simulation & Market Concentration Analyzer

Antitrust Economics, Streamlit & Google Cloud BigQuery

• Objective: Automate Phase I antitrust compliance and market concentration screening based on the U.S. DOJ Horizontal Merger Guidelines.
• Methodology: Developed a high-performance Python package (merger-sim) utilizing strict Pydantic data validation to calculate Pre/Post-Merger Herfindahl-Hirschman Index (HHI) metrics. Engineered a native SQL pipeline to query cloud data warehouses directly.
• Delivery: Built an institutional-grade Streamlit web dashboard featuring interactive parameter controls, dynamic Plotly dial gauges mapping regulatory thresholds, and stacked waterfall market share visualizations.

⚡ 3. Alpha Signal Terminal: Equity Impact Predictor

Machine Learning, NLP & Explainable AI (XAI)

* **Objective:** Forecast short-term Cumulative Abnormal Returns (CAR) on equities by quantifying unstructured market news and macroeconomic catalysts. * **Methodology:** Built an end-to-end Python ETL pipeline feeding a **Zero-Shot NLP Classifier (`BART-Large-MNLI`)** to score thematic sentiment. Trained a **LightGBM** regressor on historical feature matrices (sentiment, 30D momentum, volatility drag). * **Delivery:** Developed an institutional-grade Streamlit web application featuring **SHAP value visualization** (Explainable AI), dynamic event-study trajectory modeling, and live feature-drift governance metrics.

📊 4. Retail Media Campaign & Predictive CLV Engine

Predictive Analytics & Cloud Data Warehousing

• Objective: Extract and transform volatile transactional data to forecast long-term customer monetization paths for strategic marketing optimization.
• Methodology: Developed Python scripts to clean and aggregate purchase events into Google Cloud BigQuery. Engineered mathematical tracking modules utilizing BG/NBD and Gamma-Gamma probabilistic models for 12-month Customer Lifetime Value (CLV) forecasts.
• Business Impact: Delivered high-trust, decision-ready datasets enabling Finance and Marketing teams to evaluate cohort profitability and optimize ad spend ROI.

🤖 5. Automated Invoice Ledger Parser & LLM Pipeline

Generative AI & Unstructured Data Parsing

• Objective: Deploy NLP models to automate the extraction of complex accounting metadata from volatile, unstructured financial ledger layouts.
• Methodology: Engineered a custom text parser with token-aware sliding-window segmentation. Forced LLM completions into rigid schema patterns using dynamic JSON constraints.

🧪 6. Pricing A/B Testing & Market Optimization Simulator

Causal Inference & Experiment Design

• Objective: Design a statistical simulation environment to test dynamic price variations and measure authentic causal treatment effects.
• Methodology: Designed an A/B testing framework in Python using advanced econometric estimation loops and causal inference algorithms to map consumer demand variations while isolating external noise.

---

🛠️ Core Technical Competencies

Domain	Tools & Methodologies
Data Analysis & Econometrics	Python (Pandas, SciPy), Causal Inference, Antitrust Simulation (HHI), A/B Testing
Data Infrastructure & BI	Standard SQL, Google BigQuery, SQLite, Power BI, Streamlit
Data Science & ML	Predictive Modeling (LightGBM/XGBoost), Natural Language Processing (NLP)
Data Engineering	CI/CD (GitHub Actions), Docker, ETL/ELT Pipelines, Relational Schema Design

"Transforming abstract operational logs into definitive, data-driven business strategy."