Designing a Big Data Analytic Tool for Predicting Girl Child Learner Drop Out in the Eastern Cape Province of South Africa

The benefit of education in the socio-political development of a country cannot be underestimated. Research has shown a strong correlation between education and socio-economic factors. In South Africa, approximately, 60% of young South Africans drop out of school, without finishing Grade 12 and obtaining their matric certificate. This can be attributed to various factors ranked by the study findings such as pregnancy, drug and substance abuse, socio-economic factors, and rape. It is thus important to address these challenges early. In essence, it is crucial to determine the factors that expose the girl-child to the risks of dropping out of school. In this paper, we present the design and development of a tool to predict the likelihood that a learner would drop out of school. The tool makes use of the risk factors determined through a qualitative approach, through interviews, focus group discussions and workshops with principals, teachers, and the School Governing Body. Pregnancy was highly ranked as the main contributing factor for learners dropout rate. Using the tool, we predicted accurately the likelihood of a learner being a dropout, by providing 3 levels of ranking, high risk, medium risk, and low risk. In addition, the tool also provided insights into the relationship between the student’s living arrangements, the distance they travel to school and the financial standpoint of the parents. Developing analytic tool using data from remote areas enable smart city planners to consider disadvantaged communities. Smart schools may be difficult to achieve if dropout issues are not addressed.