ABSTRACT
Statistical learning (SL) is naturally a multidisciplinary field, bridging technical concepts from statistics, math, and computer science with specialized application domain knowledge from fields as diverse as health, industry, sports, and social sciences. Yet, SL courses tend to follow the traditional methodology from technical sciences, with a strong emphasis on lectures. Furthermore, given the role of information technologies (ITs) in SL, students are expected to present above-average IT skills. In developing countries, skills such as computational thinking (CT), in particular computer programming, are often only available in university-level technical courses, which only a small fraction of the population can attend. These requirements combine to severely restrict access to SL courses and substantially reduce the benefits derived from its multidisciplinary nature.
In this chapter, we describe a learning approach to leverage these multidisciplinary benefits. This approach is based on two principles: active learning and IT-agnostic educational objects. The first principle seeks to promote autonomy, with students organized in groups to present tutorials and develop projects. More importantly, these initiatives effectively help integrate students and bring them closer to the community and the market, with a portfolio that can be used in meetups, events, and/or job interviews. The second one seeks to include students from nontechnical fields, leveraging tools that reduce the need for an IT background. Specifically, we combine graphical user interface applications such as Spreadsheets and Orange3 with command-line interface Python libraries such as Pandas and Scikit-learn. More importantly, no educational object requires that students understand even basic computer programming concepts such as structured programming.
Initial results from a pilot project in Brazil indicate the potential of this approach with high participation from students without a strong IT background. In addition, a growing set of educational object repositories are increasing the social impact of the project.
