MS Management Information Systems · analytics, optimisation & AI

Real data, turned into decisions you can play with.

Every project here starts with a messy real-world dataset and ends in a live, interactive model. Drag a slider, change an assumption, and watch the answer change.

Explore the projects ↓ ▶ Jump into a model
0
projects
0
interactive models
0
real datasets
0
records analysed
Selected work

Projects

Two interactive portfolios, a live behavioural-research platform and an enterprise-systems analysis.

Animated map of U.S. adult obesity by state7 interactive models
Interactive portfolio

Healthcare Analytics & AI

CDC, NCHS and California hospital data turned into a causes-of-death race, an animated obesity map, a live SAS decision tree and more.

444 → 432hospital filings deduplicated
r = 0.83heart disease vs mortality
TableauSAS ViyaExcelPython
▶ Explore the models
Feasible region of a product-mix optimisation7 interactive models
Interactive portfolio

Managerial Decision Making

Optimisation and decision models that re-solve live in the browser: product mix, shipping, staffing, investing, and decisions under uncertainty.

13 / 13Excel Solver answers reproduced
$199,600optimal product-mix profit
Excel SolverLinear programmingRegressionDecision analysis
▶ Explore the models
Model pipeline diagram for the reliance studyLive study
Research prototype · ongoing

Reliance Under Disruption

Does trust in AI advice depend on where its errors fall? A 30-week warehouse simulation with a calibrated LightGBM stockout-risk model, deployed as a live study.

0.897model AUROC
30 weekssimulated per participant
PythonLightGBMBehavioural researchRender + Postgres
⚠️ Live research study. Starting a session adds a real response to the study data. Please take part only if you mean to.
Numeracy screener answers

The study opens with three numeracy questions (Schwartz et al., 1997); two of three correct are needed to continue.

  1. A fair six-sided die is rolled 1,000 times. How many times would it come up even (2, 4 or 6)? 500
  2. A lottery prize has a 1% chance of winning. If 1,000 people each buy one ticket, how many win? 10
  3. A sweepstakes car has a 1 in 1,000 chance of winning. What percent of tickets win? 0.1 (type 0.1, meaning 0.1%)
▶ Open the live studyRead the project
Order-to-cash customer concentration chartSAP S/4HANA
Enterprise systems analysis

Intelligent Enterprise Systems

SQL analysis of SAP S/4HANA sales orders, regression modelling, and end-to-end process documentation across order-to-cash, procure-to-pay, production and finance.

708sales orders analysed
5business process cycles
SAP S/4HANASQLRegressionERPsim
Read the project
Hands-on

Fourteen models you can play with

Each one runs on real data and responds as you change it. Hover to pause, click to open.

Approach

How every project is built

Find real data

Public health records, hospital filings, survey data and ERP transactions, never toy examples.

Clean it honestly

Duplicates, gaps and suppressed values are handled explicitly and documented, not hidden.

Model the question

Regression, optimisation, decision trees and clustering, chosen for the decision at hand.

Verify, then share

Every headline number is recomputed from the source, then published as something people can explore.

Python · pandasSQLExcel & SolverTableauSAS ViyaSAP S/4HANALightGBMJavaScriptData visualisationGit & GitHub

Curious how a number was reached?

Every model shows its inputs, and every project links to its data and method. Open one and change an assumption.

▶ Try a model Connect on LinkedIn GitHub