Real projects delivered across academia, finance, telecommunications, the NGO sector and government. Each card below records the methods used and the outcome measured. Full written case studies are available on request after a short discovery call.
DissertationFull quantitative PhD analysis in STATA covering five countries across ten years of panel data, using fixed and random effects estimation, Hausman specification testing, and instrumental variables to address endogeneity in the adoption measure.
Business IntelligenceA complete Power BI delivery running from the SQL warehouse through a governed semantic model to an executive dashboard, tracking more than 200 SKUs across eight distribution centres with daily refresh and alerting on stockout risk.
Predictive MLA Python classification pipeline combining Random Forest and XGBoost, validated at 91 percent accuracy on held out data, with SHAP attribution surfaced back to the retention team through an embedded Looker Studio dashboard.
Survey AnalysisLikert scale responses from 2,400 employees across fifteen member organisations, analysed through exploratory and confirmatory factor analysis, Cronbach reliability testing, and multilevel modelling that separated individual effects from organisational ones.
ETL & AutomationA Python and Excel pipeline that reads core banking data and generates branded PDF reports for twelve branches, complete with portfolio quality charts, variance commentary, and automatic distribution to every branch manager.
StrategyA data audit spanning six departments, delivered with a tool selection matrix, a governance framework covering ownership and access, and a three year roadmap that the agency went on to adopt as its official digital plan.