Data Science/Analytics

BSc Statistics (Data Science Track), Kwame Nkrumah University of Science and Technology, Ghana, 2025.

I'm strongest where statistical rigor meets stakeholder communication, the kind of work where a correlation coefficient has to become a one-page recommendation an administrator will actually use.

Consolidated five unlinked datasets spanning 63 years (1961-2023) from FAOStat, the Ghana Statistical Service, and World Bank sources. Built three live Power BI dashboards and ran correlation and regression analysis in Python to surface a structural constraint in cocoa export growth: area harvested correlates at r=0.16 with time, against r>0.85 for every food crop in the dataset. Delivered formal recommendations to Amalitech supervisors. Built with Abigail Afful.

Python pandas scipy statsmodels Power BI DAX

Compared ARIMA, ETS, and Prophet on 63 years of World Bank GDP data for Ghana. Prophet beat ARIMA by 27% and ETS by 37% on RMSE, largely because it handled structural breaks like the 1983 economic crisis better than the alternatives. Full stationarity testing, ACF/PACF analysis, and Ljung-Box residual diagnostics documented in R Markdown.

R forecast prophet tseries R Markdown

📧 samuelasteragondeze@gmail.com 💼 linkedin.com/in/samuel-agondeze-kisoke