TL;DR
This paper details the authors' experience pivoting a Text and Data Mining (TDM) course online during the COVID-19 pandemic. They share lessons learned from two pilot runs, highlighting the effectiveness of pair programming, digital badges, and continuous feedback in fostering student engagement and learning in a virtual environment.
What Problem Does This Paper Address?
The authors tackle the challenge of rapidly transitioning a hands-on, interdisciplinary Text and Data Mining (TDM) course from an in-person format to an online setting due to the COVID-19 pandemic. The core problem is how to maintain and enhance student learning, community building, and engagement in a virtual environment, especially for students with diverse backgrounds and limited prior programming experience. They aim to replicate the collaborative and social aspects of traditional classrooms online, which are often lost in pre-recorded lectures.
Methods
The authors employed several pedagogical strategies to adapt their TDM course for online delivery:
- Pair Programming: Students were randomly paired and worked together on computational tasks in breakout rooms. This aimed to foster collaboration, peer learning, and social interaction.
- Digital Badges: The course material was restructured into a modular format using digital badges. Each badge represented a "threshold concept" or core skill, providing a clear learning journey and signposting takeaways. This approach was inspired by gamification principles and aimed for flexibility and learner empowerment.
- Relentless Feedback: Continuous feedback was collected through short surveys after each session and at the end of the course. This allowed for rapid adaptation of teaching methods and content based on student input.
- Virtual Learning Environment (VLE) and Tools: They utilized the university's VLE (Learn), Blackboard Collaborate for virtual sessions (including breakout rooms, chat, and screen sharing), and Noteable (a Jupyter Notebook platform) for delivering code-along exercises and practical tasks. GitHub was used for distributing course materials.
- Hybrid Teaching Model (Pilot 2): While the initial pivot was fully online, the second pilot explored a hybrid approach, offering recorded lectures and flexibility for students joining from different time zones.
Datasets
The paper mentions the use of two specific datasets for demonstration purposes:
- Medical History of British India: Provided by the National Library of Scotland.
- Inaugural Addresses of all American Presidents (1789-2017): Used to illustrate text analysis techniques.
The authors did not introduce new datasets or benchmarks; rather, they used existing, publicly available datasets to illustrate TDM concepts.
Key Insights and Ideas
- The Power of Pair Programming Online: Despite initial unfamiliarity for many students, pair programming in breakout rooms proved to be a highly effective method for both learning and social interaction, fostering a sense of community.
- Digital Badges as a Structuring Tool: Restructuring the course content into digital badges provided a clear, modular learning path, empowering students and offering flexibility, which was particularly beneficial in a dynamic online environment.
- Relentless Feedback is Crucial: Continuous, rapid feedback loops were essential for adapting the course on the fly, addressing student confusion, and improving the learning experience. This flexibility was key to maintaining engagement.
- The Importance of a Three-Person Teaching Team: For live online teaching, having multiple instructors to manage technical issues, answer questions, and facilitate discussions simultaneously was invaluable and made the experience manageable.
- Bridging the Gap Between Theory and Practice: Integrating real-world examples from the instructors' own research helped students understand the practical applications of TDM techniques.
- The "Oddness" of Online Teaching: Even with successful strategies, the authors acknowledge the inherent challenge of not seeing student faces or body language, making it difficult to gauge engagement and enjoyment.
Key Results
- High Participant Retention: The authors were pleased with the high retention rate (at least two-thirds of students completing the course) across both pilots, attributing it in part to the relentless feedback and adaptive teaching.
- Positive Student Feedback: Students reported positive experiences, highlighting the "humour and playfulness of the examples," the effectiveness of pair learning, and the utility of Jupiter notebooks. One student even adopted the computational notebook approach for their own teaching.
- Successful Transition: The course successfully transitioned to an online format, equipping participants with initial TDM skills despite varying levels of prior experience.
- Technological Success: The authors noted that the technology used (Blackboard Collaborate, Noteable) largely worked well, which was a significant "ally" in their efforts.
Limitations and Future Work
- Gauging Engagement: A significant limitation acknowledged is the difficulty in assessing student engagement and understanding when cameras are off and participants are not physically present.
- Technical Troubleshooting: While the technology generally worked, initial sessions were impacted by technical issues, highlighting the need for robust support and attention to accessibility.
- Future Work: The authors plan to experiment with asynchronous delivery for students in different time zones, scaling the course to larger groups, and exploring writer-retreat style setups. They also intend to pivot the course back to on-campus teaching when feasible and are pleased that the course will be part of the postgraduate program at EFI. They also suggest that virtual environments will play a significant role in complementing traditional learning post-pandemic.
Summary updated