AGHEI helps ministries, agencies and NGOs decide with numbers they can defend.
Each engagement is scoped as a defined deliverable with a method note, a timeline and a hand-over package, so your team can reproduce the results after we leave.
Baseline and endline studies, M&E frameworks and impact evaluations designed around the questions funders actually ask.
Costing, cost-effectiveness, budget impact and financing-gap analysis delivered as transparent models you own.
From data audits to fair, explainable predictive models and dashboards your analysts can maintain.
The same discipline applies to a two-week costing exercise and a twelve-month impact evaluation. Every number traces back to a documented source and a line of code.
We start from the decision the evidence must support, the audience and the constraints, then write the method note before any data is touched.
Data audit, quality checks and a shared data dictionary. Where primary data is needed, instruments are piloted in French and English.
Econometric, survival, simulation or machine-learning models chosen for the question, always reported with their uncertainty and their limits.
Reports, briefs and dashboards in both languages, plus code and a training session so your team owns the result.
Most assignments run remotely with scheduled field missions. We hold Canadian and Congolese nationality, which simplifies contracting as an international or a national consultant.
Health-system monitoring, evaluation support and economic evaluation for provincial agencies, hospitals, research networks and health charities. Eligible for Canadian supplier registrations and standing offers.
Programme evaluations, costing and data systems for ministries, UN agencies and international NGOs, in French, with academic ties at HEC Kinshasa and the Université de Lubumbashi. A Kinshasa entity is being registered.
Quantitative economist and statistician, eighteen years of applied econometrics and machine learning, now applied to health systems.
John holds a PhD in Economics (Financial Economics, University of Johannesburg, 2012), a Graduate Diploma in Data Science (McMaster), an M.Ed. (Brock) and a certificate in Decision Sciences from INSEAD. He is completing a second doctorate in Health at York University's School of Health Policy and Management, specialising in health systems management and health data analytics.
Over twelve years at the University of Johannesburg he taught Bayesian and financial econometrics, statistical machine learning and risk management, coordinated the master's programme in Financial Economics, and supervised to completion more than fifteen doctoral, thirty master's and five post-doctoral researchers. He now teaches quantitative methods at HEC Kinshasa and works on fair, explainable clinical decision support and on national health-system monitoring in Canada.
He founded the Analytics Research Group, an independent think tank for quantitative modelling, whose teaching material and publication record now live under AGHEI.
A selection from more than seventy peer-reviewed articles, two edited volumes and three book chapters. The full list is on Google Scholar.
The Analytics Research Group began as a think tank for quantitative finance modelling. Its short courses continue under AGHEI and are adapted for health-sector analysts.
From data import to reproducible reports: regression, survival analysis and visualisation for programme teams.
Pandas, scikit-learn and dashboards, with a module on fairness checks for predictive models.
The original ARG lecture series, kept for finance students and useful for health-financing risk work.
Companion website
J. W. Muteba Mwamba and E. Niyonsaba Sebigunda · 2nd edition, 2026 · 24 chapters, 7 appendices, 892 pages · in French
This page holds what the printed book does not: the software manuals, the sector fact sheets for the final project, and the scripts and data for every chapter. Exercise solutions not marked with a star and post-print corrections will be published here.
Three manuals written to accompany the method chapters; they refer to the book by chapter number.
Gretl : guide de prise en mainPDF, 43 pages. Menus, hansl scripts and output, with the datasets shipped with Gretl. JMulTi : analyse VAR et cointégrationPDF, 29 pages. Unit roots, VAR, VECM, diagnostics and structural analysis. R : guide d'utilisationPDF, 43 pages. Packages, code and a chapter-by-chapter package list.Eight sheets for the chapter 24 final project: health, education, microfinance, employment, agriculture, conflict, social protection, informal economy. Each gives typical questions, data sources, the book's methods adapted, and pitfalls.
Fiches sectoriellesPDF, 4 pages.Chapter scripts (Python, R, Gretl) with their output and the datasets they read, in the folder structure the scripts expect. Run them from the root of the unzipped folder.
Code et donnéesZIP, about 2 MB. LISEZMOI_donnees.md describes each dataset and its source.The CIT_2020_CUP and CIT_2024_CUP files are not redistributed; download them from github.com/rdpackages-replication and place them in the data/ folder.
Solutions to the exercises marked with a star are in appendix D of the book. Solutions to the other exercises, corrections reported after printing and updates to cited sources (databases, software versions) will be published on this page.
Questions about the book: use the contact form below and choose "Training".
Short, practical pieces on monitoring health systems and evaluating reforms. New notes are added as projects close.
What 6,757 Canadian headlines, sentiment models and Granger-causality networks reveal about workforce, affordability and access pressures, and how a ministry could use the same pipeline as an early-warning tool.
How survey data, qualitative coding and text analysis were combined to measure the effects of the Bachelor-Master-Doctorate reform, and what transfers to health-sector reforms.
We reply within two working days with a short scoping note: what we would measure, how long it takes and what it costs.