Selected projects
I use analytics to investigate workforce behaviour, predict risk, forecast demand and support evidence-based organisational decisions.
This portfolio is intentionally analytical: each project shows the business question, method, evidence, limitations and decision value. Full assessed submissions are not reproduced publicly; each button opens a dedicated one-page analytical summary.
Predictive analytics · Retail revenue
Forecasting Global Apple Product Sales
I compared Linear Regression, Random Forest and XGBoost on 11,500 product-sales observations, using feature engineering, leakage control and model diagnostics to estimate revenue.
PythonXGBoostR² ≈ 0.998Feature importance
HR analytics · Explainable AI
Employee Attrition Prediction Research
I designed a comparative study of Logistic Regression, Random Forest, Gradient Boosting and SVM, balancing predictive performance with transparency for responsible HR decision-making.
SHAPAUC-ROCCross-validationResearch design
Supply chain analytics · Classification
Predicting Global Supply Chain Disruptions
I developed a leakage-aware classification workflow using 10,000 shipment records and compared Logistic Regression with Random Forest for disruption detection.
94.15% accuracyPrecision 1.00Recall analysisRandom Forest
Business statistics · Credit risk
Geographic Drivers of Credit Default
I used Chi-square and Kruskal-Wallis testing on more than 2,000 credit records to examine borrower default patterns and geographic differences in overdue exposure.
SPSSχ² = 224.246Cramer’s V = 0.331p < 0.001
Data visualisation · Health analytics
Predictors of Anxiety Using R
I analysed 11,000 observations to examine the relationships between anxiety, stress, sleep, diet, physical activity and psychosocial factors through visualisation, correlation and regression.
RR² = 0.666r = 0.668Regression
Group consultancy · Forecasting & optimisation
Retail Sales Forecasting and Profit Optimisation
As part of Team Value Predictors, I contributed to an end-to-end retail analytics proof of concept combining SARIMA and ETS forecasting, regression, classification, K-Means segmentation and market-basket analysis to improve inventory and margin decisions.
SARIMA MAPE 18.97%MASE 0.43Decision TreeK-Means20% safety stock
Strategy · Operations simulation
Molniya Automotive Simulation
I evaluated four rounds of strategic decision-making across finance, production, market share, workforce planning and team governance, including scenario testing for a new-model launch.
£1.281bn profit1.85% market shareScenario testingOperations