Moodoo the Tracker
Spatial Analytics for Characterising Classroom Pedagogies
Our aspirations for this project are to improve the quality of student learning in collaborative, group and laboratory learning environments. In this context, the teacher’s ability to create an appropriate curriculum for such environments and navigate the learning space with confidence and efficiency is crucial to the students’ learning experiences. This project focuses on Spatial Pedagogy and the teachers’ spatial behaviours in the classroom.
Teachers’ spatial behaviours in the classroom can strongly influence students’ engagement, motivation and other behaviours that can shape their learning. However, classroom teaching is ephemeral, and has largely remained opaque to computational analysis. Inspired on the notion of Spatial Pedagogy, we are delevoping a system (‘Moodoo’) that automatically tracks and models how teachers make use of the classroom space by analysing indoor positioning traces. To investigate the potential of the system, we conducted an authentic study with seven educators enacting three distinct learning designs to more than 200 undergraduate students in the context of science education.
Contrasting spatial pedagogical approaches. Left: a teacher focusing on certain students during a class. Centre: a balanced teacher who spread her time across various groups of students. Right: a second teacher mostly walking around the classroom, supervising.
Moodoo automatically extracts spatial metrics (e.g. teacher-student ratios, frequency of visits to students’ personal spaces, presence in classroom spaces of interest, index of dispersion and entropy), mapping from low-level educator’s positioning data to higher-order spatial constructs. We illustrate how these spatial metrics can be used to generate a deeper understanding of how the pedagogical commitments embedded in the learning design, and personal teaching strategies, are reflected in the ways educators use the learning space to provide support to students.
This is a mutli-center multidisciplinary project spearheaded by Roberto Martinez-Maldonado (Monash University, Australia) in collaboration with Vanessa Echeverria (Escuela Superior Politécnica del Litoral, Ecuador), Katerina Mangaroska (Norweigan University of Science and Technology, Norway), Antonette Shibani, Gloria Fernandez-Nieto, Jurgen Schulte and Simon Buckingham Shum (University of Technology Sydney, Australia).
Digital Analytics for Classroom Proxemics
Classroom Proxemics looks into to how teachers and students use classroom space, and how this and the spatial class room design impacts on learning and teaching. In particular, we aim to look into the development of an automated feedback that helps teachers identify salient patterns in their use of the classroom space. Here, we make use of indoor positioning data from four teachers, analyse the data and use it as evidence to compare in a visual representation three distinct learning designs enacted in a physics laboratory classroom. Teachers can make effective use of such visualisations, to gain insight into their classroom practice through reflections on visualisations of positioning data, both their own and that of peers; and identifying the types of indicator (operationalised as analytical metrics) that foreground the most useful information for them to gain insight into their practice.
Three example maps of sessions 4, 7 and 11 (taught by the same teachers, T2 and T5) in which territoriality aspects were highlighted by teachers. T2’s data points are shown in blue and T5’s in orange.
Three annotated maps in which specific aspects associated with classroom proxemics were explained by teachers based on the three different learning designs (LD1-3).
This is a mutli-center multidisciplinary project spearheaded by Roberto Martinez-Maldonado (Monash University, Australia) and Jurgen Schulte (University of Technology Sydney, Australia), Vanessa Echeverria, (Carnegie Mellon University, USA), Yuveena Gopalan and Simon Buckingham Shum (University of Technology Sydney, Australia).
Learning Analytics for Individualized Student Support Actions
All the while I was creating these fantastic truly authentic learning environments that turned out to be so much fun for everyone involved, the just-in-time individualized student support became permanent companion and integral part in all my courses (Progress Analytics 0.1). The net of just-in-time data collection swept ever wider and deeper and the ‘automation’ of just-in-time student feedback and supportive advise pushed me into the wonderful entrails of Learning Analytics. The Learning Analytics I am interested in lives at the coalface front of direct, actionable student learning support rather than the bird’s eye institutional view. Although, what I’d really like to do is to create an easy accessible, seamless link between the two (Future Projects).
Schulte J (2015)
Gamification of learning within a problem solving context – Individualised learning, adaptive testing, confidence and proficiency self-ratings
UTS First Year Experience Forum 23 September 2015.
Pardo A, Schulte J
Scaling Instructor-driven Personal Support Actions (Workshop)
ALASI 2015: Australian Learning Analytics Summer Institute. 26 – 27 Nov 2015, Sydney.
Schulte J (2016)
Experiences with an Adaptive Learning Product – A Learning Analytics Perspective
The NSW Learning Analytics Working Group Meetup. 3 February 2016, Sydney.
Pardo A, Martínez-Maldonado R, Buckingham Shum S, Simon McIntyre S, Gasevic D, Siemens G, Schulte J (2017)
Connecting Data with Student Support Actions – A hands-on Tutorial.
LAK17 Learning Analytics and Knowledge Conference.March 13-17 2017, Vancouver, Canada.
Schulte J, Buckingham Shum S, Martinez Maldonado R (2017)
Large Scale Predictive Process Mining and Analytics of University Degree Courses.
LAK17 Learning Analytics and Knowledge Conference. March 13-17 2017, Vancouver, Canada.
Course Pathways: Making informed choices
Jurgen Schulte, Martinez-Maldonado, Simon Buckingham Shum
The aspiration of this project is to uncover statistically significant patterns in students’ course pathway choices with the help of data mining and with this information to derive individual student course-longitudinal ‘health’ indicators.
This is done with a view to provide support units, course and subject coordinators with more longitudinal focused indicators that may be used in student personal support actions (on-demand or just-in-time individualised student support). The indicators may also help to support the streamlining of course and subject content. The more students can be informed about what it would take for them (individually) to master future subject and stages in their course, the better their study experience will be and a higher overall student retention rate may be achieved. This project is a collaboration with the UTS Connected Intelligence Centre. The proof of principle realisation of the statistical analysis and projection of student data is done in R. Pre-processing data mining of 3 million UTS student enrollments in over 1,200 degree programs is done with Rapid Miner and KNIME. Analytics extraction of longitudinal focused student pathways, performance indicators and visualisation is realised with R.
The first public demonstration of the pathways dashboard tool has been at the UTS Teaching & Learning Forum Showcase in November 2016. In the further development of this tools we found evidence that process mining of student course pathways and student performance rapidly degrades performance as the number of course pathway and students increase (LAK 2017). We moved away from the process mining and developed a smart large scale data filtering approach of exceptional performance which now allows us to interrogate arbitrary student course pathways in almost real-time.
Now that we have rapid access to the course pathway performances, we’ll be focusing on the user interface and dissemination of this tool.