Learning Analytics: Serving the Learning Process Design and Optimization
Yanyan Li, Haogang Bao and Chang Xu
Abstract Data growth in the information era is changing commercial and scientific research. In educational settings, a key question to address is how to effectively use the massive and complex data to serve the teaching and learning optimization. Therefore, as an emerging data analysis technology, learning analytics increasingly draws more attention. This paper proposes a process model of learning analytics, reviews the research and challenges of multi-source educational data collection and storage, generalizes typical data analysis approaches, and elaborates on how to align learning analytics with pedagogical and organizational goals.
Keywords Learning analytics Process model Educational data collection Pedagogical context
The proliferation of diverse learning environments, such as learning management systems (LMS), online learning communities, personal learning environments (PLE), and adaptive learning systems, produces impressive amounts of data. Different types of data relating to learners’ learning process can be tracked and stored, which makes it possible to identify the preferences of the students, refine teaching methods to better fit students’ needs, and provide empirical evidence for educational decision-making. Nevertheless, exploitation of such available educational data is still rare. In educational settings, a key issue to address is how to
Y. Li (H) • H. Bao • C. Xu
Beijing Normal University, Beijing, China
F.-Q. Lai and J.D. Lehman (eds.), Learning and Knowledge Analytics
in Open Education, DOI 10.1007/978-3-319-38956-1_4
effectively use the massive and complex data to serve the teaching and learning optimization.
Learning analytics (LA) is an emerging field in which sophisticated analytic tools are used to improve learning and education. Different from other similar fields, like business intelligence, academic analytics, and educational data mining, learning analytics aims at serving the design and optimization of the learning process. While some studies have been carried out on LA, other issues that potentially affect the acceptance and impact of LA, like the compatibility of datasets, privacy, and interpretation of results, still remain to be explored. The implementation of LA in the learning process also must be carefully crafted in order to be successful and beneficial (Greller and Drachsler 2012).
This necessity motivated researchers to identify critical LA factors or dimensions which needs to be reckoned with to ensure an appropriate exploitation of LA in educational settings. Elias proposed an ongoing three-phase cycle that aims at the continual improvement of learning and teaching. The three phases comprise data gathering, information processing, and knowledge application (Elias 2011). With a general morphological analysis approach, Greller and Drachsler (2012) outlined a generic design framework composed of six dimensions (i.e., stakeholders, objectives, data, instruments, external constraints, and internal limitations) deduced from discussions in the emerging research community. In addition, Chatti et al. (2012) described a reference model for LA based on four dimensions, namely data and environments (what?), stakeholders (who?), objectives (why?), and methods (how?).
By summarizing the essential factors of LA, this paper proposes a process model of learning analytics from the data processing cycle perspective. The remainder of this paper is organized as follows. Section 4.2 presents the process model of learning analytics; Sect. 4.3 introduces the status and challenges of multisource data collecting and storage; Sect. 4.4 generalizes typical data analysis approaches; Sect. 4.5 elaborates on how to optimize learning and teaching, and Sect. 4.6 concludes the paper.