Development and Evaluation of an Approach for Integrating Data Science Concepts into High School STEM Curriculum

Currently STEM education evolves rapidly towards a higher integration either among separate constituents (Science, Technology, Engineering and Mathematics) or by adding the new aspects, such as social. This paper aims at extending the science dimension in STEM-driven Computer Science (CS) education by integrating (through modelling) Data Science (DS) concepts into the high school curriculum. Three types of models and modelling (conceptual, feature-based, and physical modelling) incorporated into a coherent methodology form the background of the proposed approach. Models and modelling, as well processes with data, are essential attributes of engineering education too. The core result of this paper is a novel three-layered framework outlining a series of modelling processes to support integration along with the assessment model. The latter includes the Revised Bloom's taxonomy combined with computational and scientific thinking skills. The use of this approach in the real educational setting and provided experiments show that discovered models ensure a seamless integration of the DS component into STEM-driven CS education. This approach contributes to the increased students' motivation to learn due to the interesting real-world task and active learning of engineering aspects through constructing and testing own experimental system and data processing. The approach also enforces the learner's interdisciplinary knowledge by computational, scientific, and designing skills so important for engineering activities.