nkharemwa/EduEvidence-Kenya

A four-project data science portfolio examining learning outcomes, teacher effectiveness and student risks in Kenya's education system. Methods include Causal Inference, Interpretable ML, NLP, and Deep-Learning built in R and Python for development sector audiences. EduEvidence Kenya is a data science portfolio built around a single question: What does evidence-based education look like in practice? Across four projects spanning Causal Inference, Machine Learning, Survey Data Engineering, and Deep-Learning, this portfolio demonstrates how rigorous analytical methods can be applied to real education challenges in Kenya: from identifying students at risk of dropping out, to evaluating the impact of teacher training programs, to extracting insight from messy field survey data. All projects are designed with deployment in mind. Models are interpretable, outputs are actionable and findings are documented for both technical and non-technical audiences.