In this paper, we propose a counterfactual explanations-based approach to provide an automatic and intelligent recommendation that supports student's self-regulation of learning in a data-driven manner, aiming to improve their performance in courses. Existing work under the fields of learning analytics and AI in education predict students' performance and use the prediction outcome as feedback without explaining the reasons behind the prediction. Our proposed approach developed an algorithm that explains the root causes behind student's performance decline and generates data-driven recommendations for action. The effectiveness of the proposed predictive model that constitutes the intelligent recommendations is evaluated, with results demonstrating high accuracy.

The advancements in learning analytics and artificial intelligence have shown potential to transform traditional modalities of education. One such advancement relates to the use of educational data to track students’ knowledge state [1] . In the field of Artificial Intelligence in Education knowledge tracing is a well-established area where a machine models the students’ knowledge as they interact with coursework. Effective modeling of student knowledge can have a high impact on the provision of adaptive learning. In fact, lately, research on knowledge tracing is intensifying with a particular focus on the utilisation of new machine learning algorithms for modelling the students’ knowledge levels and for the prediction of performance on future tasks and assessment questions [2] . In the case of question-level assessment, knowledge tracing provides an interpretation of the learner’s current knowledge level and models their mastery of the skill or knowledge component to which future questions are related [3] .

There has been significant efforts in studying collaborative and social learning using aggregate networks. Such efforts have demonstrated the worth of the approach by providing insights about the interactions, student and teacher roles, and predictability of performance. However, using an aggregated network discounts the fine resolution of temporal interactions. By doing so, we might overlook the regularities/irregularities of students' interactions, the process of learning regulation, and how and when different actors influence each other. Thus, compressing a complex temporal process such as learning may be oversimplifying and reductionist. Through a temporal network analysis of 54 students interactions (in total 3134 interactions) in an online medical education course, this study contributes with a methodological approach to building, visualizing and quantitatively analyzing temporal networks, that could help educational practitioners understand important temporal aspects of collaborative learning that might need attention and action. Furthermore, the analysis conducted emphasize the importance of considering the time characteristics of the data that should be used when attempting to, for instance, implement early predictions of performance and early detection of students and groups that need support and attention.

The association between mathematics and programming in an educational context is not new. Today, programming has been introduced into curricula worldwide for younger children. In the Swedish case, primary school teachers are expected to integrate programming in mathematics education from autumn 2018. However, Swedish teachers' knowledge of programming and programming didactics is limited. Meanwhile, there is little research on K-9 programming education. This has led to the dilemma that the mathematics teachers have limited support in didactic knowledge and good examples. This study reports on a teacher professional development project in programming. More specifically, teachers used Lesson Study to plan, execute, and evaluate lessons that integrated programming into various school subjects in elementary school. This study analyzed the didactic strategies developed in 10 lesson studies, as well as mapped the opportunities and challenges of pupils' learning in the mathematics subject. The result was the identification of three didactic strategies, which were analog programming, robot programming and block programming, as well as 11 didactic methods applied within these strategies. The paper contributes with examples of the didactic methods that teachers have developed and evaluated using lesson study. The paper further provides insights on how teachers can take progression into account by applying the three didactic strategies. At last but not least, the study shows a great need for teachers to develop computational thinking abilities.

The master thesis is the last formal step in most universities around the world. However, all students do not finish their master thesis. Thus, it is reasonable to assume that the non-completion of the master thesis should be viewed as a substantial problem that requires serious attention and proactive planning. This learning analytics study aims to understand better factors that influence completion and non-completion of master thesis projects. More specifically, we ask: which student and supervisor factors influence completion and non-completion of master thesis? Can we predict completion and non-completion of master thesis using such variables in order to optimise the matching of supervisors and students? To answer the research questions, we extracted data about supervisors and students from two thesis management systems which record large amounts of data related to the thesis process. The sample used was 755 master thesis projects supervised by 109 teachers. By applying traditional statistical methods (descriptive statistics, correlation tests and independent sample t-tests), as well as machine learning algorithms, we identify five central factors that can accurately predict master thesis completion and non-completion. Besides the identified predictors that explain master thesis completion and non-completion, this study contributes to demonstrating how educational data and learning analytics can produce actionable data-driven insights. In this case, insights that can be utilised to inform and optimise how supervisors and students are matched and to stimulate targeted training and capacity building of supervisors.

Effective collaborative learning is rarely a spontaneous phenomenon. In fact, it requires that a set of conditions are met. Among these central conditions are group formation, size and interaction dynamics. While previous research has demonstrated that size might have detrimental effects on collaborative learning, few have examined how social dynamics develop depending on group size. This learning analytics paper reports on a study that asks: How is group size affecting social dynamics and performance of collaborating students? In contrast to previous research that was mainly qualitative and assessed a limited sample size, our study included 23,979 interactions from 20 courses, 114 groups and 974 students and the group size ranged from 7 to 15 in the context of online problem-based learning. To capture the social dynamics, we applied social network analysis for the study of how group size affects collaborative learning. In general, we conclude that larger groups are associated with decreased performance of individual students, poorer and less diverse social interactions. A high group size led to a less cohesive group, with less efficient communication and less information exchange among members. Large groups may facilitate isolation and inactivity of some students, which is contrary to what collaborative learning is about.