A long-form analytics article on why KPIs fail when the denominator silently changes. Covers denominator contracts, eligibility vs observed vs attempted populations, drift/coverage failure modes, mix shifts, time alignment, and a practical audit checklist to design decision-safe ratio metrics that don’t lie.

Lord's paradox

Mixed effect modeling with R tutorials

Team Beta Group Project on UC Berkeley Admissions Analysis

Reproducible code for our paper "A Differential Measure of the Strength of Causation"

Higher Diploma in Science in Computing (Data Analytics) - Programme Module: Fundamentals of Data Analysis (COMP07084)

SaaS onboarding A/B test analysis using Python, scipy, and statsmodels. Covers experiment design, statistical testing, Simpson's Paradox discovery, and sequential testing pitfalls.

Address the admission rate of University of California, Berkeley admission data to determine if their admission process was gender-biased.

A new but puzzling condition for agglomeration in Bayesian confirmation.

Presenting the Simpson's Paradox on Covid-19 data

Task assessment

Quantitative investigation of the so-called Alan Author Effect in Bayesian confirmation theory. Results for Simpson's Paradox are also provided.

an analysis of Simpson's paradox example dataset

📊 Identify KPI miscalculations by focusing on the right denominator to improve decision-making and prevent business illusions.