Built an ARIMA/ETS forecasting model for US Personal Consumption Expenditure achieving MAPE under 0.2% — well within management reporting standards. In parallel, performed NLP sentiment analysis on 2,000 hotel reviews to surface actionable operational insights.
End-to-end consulting engagement for a simulated retail client — gathered requirements, ranked 5 candidate sites using multi-criteria decision analysis (WSM/TOPSIS), identified sales drivers through OLS regression, and delivered board-level recommendations. Direct parallel to associate consultant delivery.
Designed and built an end-to-end NLP classification pipeline to detect counterfeit product reviews, achieving 94% accuracy using SVM with TF-IDF feature extraction. Full process documentation and methodology notes produced to professional delivery standard.
Designed and executed a structured research programme across 20 stakeholder interviews exploring blockchain and cryptocurrency adoption in real estate. Managed requirements, timelines, risk, and final reporting end to end — applying thematic and sentiment analysis to surface strategic recommendations.