Compare students from University of Brasília by Gender Using t-SNE Techniques

Resumo

The presence of women in technology-related courses is declining every year, reaching in 2016 less than 20\% of the total student body in the field of Computer Science in Brazil. This paper proposes to study visualization techniques to analyze and identify profile patterns of undergraduate students on courses in the computing field, comparing gender data (female and male). A quantitative analysis and dimensionality reduction technique (t-SNE) were used for the knowledge discovery and a visual data analysis process, considering the students situation at $$ (active, drop out, graduated, death and others). The visualizations revealed that the high number of students leaving without graduating is not a gender-related problem. Also the quantitative analysis shows that quotas' students are not a important information between those who leave without graduating.