Status: Development in progress
Analytics is disabled in learn.moodle.net. To enabled it would be useful for different things, mainly:
- To test our own product in a production site and identify its major flaws, usage and UX issues...
- To identify students at risk of dropping out of the MOOC
- To identify possible new features for it. E.g. bulk suggested actions like messaging
This issue has been discussed with Helen Foster and she mentioned that it is very likely that Mary Cooch is also keen to enable it. It would be great if we could enable it after the current session of the MOOC so we are 100% sure that nothing will interfere with the MOOC session.
To enable analytics is quite simple as a PHP machine learning backend is included in Moodle. There is a 2nd machine learning backend in core that requires a python package. Given the learn.moodle.net size and its superior capabilities, the python backend is the preferred option. It requires python 2.7 and the following pip package: pip install moodlemlbackend (more info in https://docs.moodle.org/36/en/Analytics#Predictions_processor)
Setup - I (David Monllao) together with Helen or Mary can go through these steps.
- Go to Site administration > Analytics > Analytics settings
- Set Default predictions processor to ''Python machine learning backend*
- Ensure that ''Models output directory* exists and is writable
- Set Time splitting methods to Quarters accumulative
- Go to Site administration > Analytics > Analytics models and, for Students at risk of dropping out, select Actions > Edit
- Enable the model
- Set Time-splitting method to Quarters accumulative
- Ensure that the python machine learning backend is the pre-selected option for Predictions processor field
- Save changes