CoTrack is a Multimodal Learning Analytics tool which offers an easy-to-use interface for multimodal data collection from collaborative learning activities.
CoTrack allows teachers/researchers to cusomize the group activity and data recording. For example, CoTrack supports two different types of group activities: first, which only involves use of collaborative text editor; second without the text editor.
CoTrack captures following types of data and provides pre-processed features
- Voice activity detection (VAD): Data indicating whether someone speaks or not.
- Speech to text: Transcript from audio data using Google speech-to-text api in real time (only supported in chrome browser).
- Writing logs: Log data of participants' writing in collaborative text editor. Activity monitoring
CoTracks offers a real-time dashboard (updates every 30 seconds) presenting a social network (who is talking after whom network) and groups' writing activities in terms of number of revisions made.
CoTrack integrates machine learning models to predict collaboration quality of a group based on speaking activity distribution, features from speech data, and writing activities. Based on predicted collaboration quality, some recommendations are provided to assist teachers in improving the siution if detected LOW.
- Backend: Django
- Database: MySQL (recommended)
- Frontend: HTML, Bootstrap5, Java Script
🚧 Under active development
- Customizing usage experience based on human-AI teaming levels such as situational and operational.
- Integrates prediction into analytics dashboard
- Storage of predicted values in database for later viewing
The following steps offer guidelines on setting up & running SEEDS app on a local machine.
git clone https://github.com/pankajchejara23/Taico-cotrack.git
cd Taico-cotrack
pip install -r ./requirement.txt
python manage.py makemigrations
python manage.py migrate --run-syncdb
The following command creates an admin user account.
python manage.py createsuperuser
Everything is set now to run the server.
python manage.py runserver
When putting CoTrack in production, make sure Etherpad is already installed and running. CoTrack uses Etherpad and might fail if Etherpad is not available. Refer to resources for instructions.