It’s been two years since Moodle HQ announced that Project Inspire, a next-gen analytics initiative, was in the making. Its pace of development has not kept in line with the Moodleverse anticipation, which is why the following recap could be worthwhile.
- Project Inspire would be led by Elizabeth Dalton, Learning Analyst for Moodle HQ. During the past year, she earned her Ph.D. in Curriculum and Instruction from the University of New Hampshire with a focus on Learning Analytics. (Congrats, Liz!)
- It would source data from Moodle users and installations around the world. The data would be provided voluntarily by registered Moodle site admins and would be “anonymized” to protect the users’ identities.
- It would involve some degree of machine learning.
By the turn of last year, not a lot was revealed, except for Dalton’s talk at MoodleMoot Australia in December. Even though it did not mention Inspire by name, the talk did a great job in highlighting the technical and pedagogical assumptions behind true learning analytics, and not those solutions “carried over” from digital marketing or business metrics.
A few months into 2017, four Courses/Discussion Forums were launched at the Moodle forums: Data Collection, Project Inspire Background, Roadmap, and Working Group. This is arguably the beating heart of Project Inspire’s progress. As with most early stage initiatives, reception and activity by users are the main predictors of timely development.
Several weeks before the launch of Moodle 3.3, documentation for the Project Inspire plugin started to detail the Inspire Plugin, a predictive tool that will offer levels of confidence for future student success according to their behavior, and even prescriptive commands in case remedial action is necessary.
Finally, the Inspire plugin, which is the most explicit outcome of the project to date, provides some limited forecasting functionality. Ideally, the predictive accuracy of the plugin will be validated by third-party examination of the code and a controlled evaluation in empirical scenarios, none of which has been announced to date. While it does offer some model evaluation functionality, the current version of the Inspire plugin does not include a feedback loop and it only uses data from the Moodle Site where it’s implemented. Whether the plugin draws from the anonymous Moodle data set of users worldwide is still unclear. If your Moodle site has no records of past Course behavior, the model has no learning set and would not be able to make predictions.
As for the future, the Inspire plugin is expected to join the Moodle core. It is not clear if the progress will allow this to happen for Moodle 3.4, scheduled for November this year. On June 1st, an unofficial statement by the CTO of Spanish Moodle Partner Inserver during Open Expo España stated that Analyse, either a new plugin or a next stage of the original Inspire plugin, will be ready for Moodle 3.4.
Install and download the Inspire plugin here. (Moodle 3.3 and PHP 7 required.) The plugin is maintained by David Monllaó, with Dalton’s help.
For information about getting involved with Project Inspire, read our guide. For technical details, you can read a preliminary overview of the Inspire API here.