In the learning environment of the project the students will be immersed within networked labs (physical and virtual) and they are tasked to solve practical problems and discover solutions by experimenting with tools, materials, and artifacts, by applying experiential learning approach which includes practical and hands-on learning methods. This lab-based learning approach enables students to directly manipulate materials, electronic components, sensors, energy and information, and provide adequate opportunities to apply their knowledge and efforts to find creative solutions for real problems with the provided laboratory equipment. This learning setting with the learning scenarios offer ample opportunities for providing learning analytics to support the acquisition of skills and holistic professional competences of engineering students and enhance their laboratory based learning experiences.
Although lab-based learning coupled with experimentation and crafting is a cornerstone for educating engineers and enabling them to gain practical skills, there is very little research and experience in terms of learning analytics, and especially multi-modal learning analytics. The main goal of Multimodal Learning Analytics (MLA) research is to extend the application of learning analytics tools and services in learning contexts which do not readily provide digital traces and learning data. Moreover, the characteristics and properties of these learning contexts cannot be described by a single source of data traces, but a combination of several modes and sources are vital in understanding these particular learning processes. In the DigiLab4U project we are developing a connected lab infrastructure with learning analytics service as an integral component. The learners within our networked lab infrastructure while doing their learning activities will generate a wide range of analogue and digital learning-related data that can be analyzed and the analytics results used to support and enhance lab-based learning experiences.
Figure 1 sketches out the data aspect of the analytics setup which is part of the learning analytics component of the networked lab infrastructure. The learning activities in these labs include interactions within the learning platform online (including the wide range of learning resources and instruction), physical interactions with the lab equipment, physical presence and movement in the lab including group work, time-spent within the lab, various sensor data from the labs, formative and summative assessment data, and interactions and events from the AR/VR equipment installed for the interconnection of the labs.
Each of these interactions are transformed into uniformed events which will be saved within a central data warehouse. The combination and correlation of this multi-modal learner data has to be aggregated and properly stored within one central Learning Record Store (LRS) and the data should retain its semantic value, and ensure that the data is available for analysis and interpretation while conforming to current data privacy regulations within the EU. The data warehouse in our scenario will be an implementation of Learning Locker with xAPI as an underlying data standard. The didactical approaches for experiential learning call upon a wide range of problem-solving, and feedback and debriefing activities. This influences the choice of data analysis methods and algorithms with which will be used for analyzing the learner data. We are planning to apply different data analysis methods (statistics and data mining methods) to explore the data and look for connections, interrelations and development within the multi-modal data. Depending on the didactical approaches and student feedback types, some of the data would be analyzed in batches, while some has to be analyzed (and potentially visualized) in real-time. Figure 2 outlines the flow from raw data towards analytics results and interpretation.
The last component of the learning analytics within these labs is the delivery, use, and interpretation of the data analytics results. The mixture of interfaces (web, physical tools, and AR/VR technology) provides several possibilities of providing analytics results, such as, learning dashboards on the learning platform, or stand-alone learning dashboard, inside the AR/VR environment, or some analytics results could be embedded within the experiments results. We will look at various analytics delivery options which again will be influenced by the didactical questions and learning goals of the stakeholders. Hence, appropriate learning analytics indicators, and visualizations will be developed and delivered with the above-mentioned learning analytics delivery approaches. The technical implementation of the described learning analytics component is a substantial step towards providing MLA tools and services in laboratory based learning scenarios to support the acquisition of skills and holistic professional competences of engineering students and enhance their laboratory based learning experiences.