Projects
Project 1: Global Health Expenditure Efficiency
Status: Complete | Tools: Python, pandas, matplotlib | Data: WHO Global Health Expenditure Database (2015–2022)
Question: Which countries achieve different patterns of health financing in terms of government spending, out-of-pocket expenditure, and health spending relative to GDP?
Key Findings:
- Large spending gap: Health expenditure per capita varies substantially across countries, with the United States spending considerably more than lower-income countries included in the analysis.
- Government financing matters: Countries differ substantially in the share of healthcare spending financed through government sources versus households.
- US spending is an outlier: The United States has a substantially higher health expenditure share of GDP than the European countries examined.
- Out-of-pocket burden: Several lower-income countries rely much more heavily on direct household payments for healthcare.
Policy Implications:
- The analysis highlights substantial differences in healthcare financing structures across countries.
- High household out-of-pocket spending can create financial barriers to healthcare access.
- Cross-country comparisons provide a useful starting point for examining health-system financing and efficiency.
View full chart | View code on GitHub
Project 2: The COVID-19 Health Financing Shock
Status: Complete | Tools: Python, pandas, matplotlib | Data: WHO Global Health Expenditure Database (2019–2021)
Question: When COVID-19 hit, how did government and household healthcare financing change across different health systems?
Key Findings:
- Different financing responses: Countries varied considerably in how public and household health expenditure changed during the pandemic.
- Household burden: Some countries experienced increases in out-of-pocket spending despite increases in overall health expenditure.
- Cross-country variation: The analysis identified different patterns of financing response across the countries examined.
- Health-system resilience: Changes in public financing provide one perspective for examining how health systems responded to the financial shock of COVID-19.
Policy Implications:
- Government financing capacity can influence how health systems absorb major shocks.
- Changes in household out-of-pocket spending are important when assessing the financial consequences of a health crisis.
- Public financing data can provide an initial basis for comparing health-system responses to major disruptions.
View full chart | View code on GitHub
Project 3: Pharmacovigilance Signal Detection — FDA FAERS Q1 2026
Status: Complete | Tools: Python, pandas, matplotlib | Data: FDA Adverse Event Reporting System (FAERS) Q1 2026
Question: What patterns can be identified in publicly available post-market adverse event reporting data?
Analysis:
- Processed a large publicly available FAERS dataset from Q1 2026.
- Examined reported drugs, adverse-event terms, and serious outcome categories.
- Used Python to clean, aggregate, analyse, and visualise pharmacovigilance data.
- Explored reporting patterns that could be used to identify areas for further safety investigation.
Important limitation: FAERS is a spontaneous reporting system. Report counts cannot by themselves establish incidence, causality, or comparative drug safety.
View full chart | View code on GitHub
Project 4: Semaglutide Budget Impact Models
Status: Complete | Tools: Python, pandas, data analysis and visualisation | Data: Public healthcare and official population sources
Question: What could be the potential five-year budget impact of semaglutide-based obesity treatment across different healthcare systems?
This project develops comparative Budget Impact Models (BIMs) for semaglutide (Wegovy) obesity management across the United Kingdom, United States, and Italy.
Model components:
- Adult population estimation
- Obesity prevalence and eligible population estimation
- Treatment uptake scenarios
- Annual treatment costs
- Healthcare cost offsets where country-specific evidence is available
- Five-year gross and net budget impact
- Price and assumption sensitivity analysis
Research approach: The models use publicly available population, epidemiological, healthcare, and pharmaceutical data. Country-specific assumptions and sources are documented in the repository to make the calculations traceable and reproducible.
View the complete Semaglutide Budget Impact Models project on GitHub
Project 5: European Food Policy & Obesity Analysis
Status: Complete | Focus: Health Economics, Public Health, Food Policy
This project examines the relationship between food-policy characteristics and adult obesity prevalence across 27 European countries using publicly available data.
The analysis combines country-level obesity prevalence data with indicators describing food-policy activity and policy instrument diversity. I used Python and OLS regression to explore cross-country associations and conducted model diagnostics, including residual and influence analysis.
Key model results:
- 27 countries included
- R² = 0.1965
- Adjusted R² = 0.1296
- F-statistic = 2.935
- Overall model p-value = 0.0724
The model explained approximately 19.7% of the variation in obesity prevalence. The overall model does not meet the conventional 5% statistical significance threshold, so the findings are presented as exploratory associations rather than evidence of a causal relationship.
The project includes regression coefficients, model diagnostics, residual analysis, visualisations, and documented data sources.
Tools: Python, pandas, NumPy, statsmodels, Matplotlib, Seaborn, Jupyter Notebook.
Current research interests: health economics, public health policy, digital health, real-world evidence, pharmacovigilance, genomics and healthcare data analytics.