Blended Learner Data Science (BLDS). Use-case for AppInventor
Analyse and visualize data from the appinventor course and the corresponding EDX server and AppInventor server data.
Python 3, Pandas, Numpy, Colorspacious, Dill, Networkx, Pillow, Python_dateutil, Python_Lenvenshtein, Matplotlib, Seaborn, SKLearn, Textstat, Jupyter, PyGraphviz
For the packages installable through pip, you may use the requirements.txt
to install them.
Visualize appinventor sessions aggregated among all students.
Visualize the number of days students have clickstream activity.
Quanitfy how students move between knowledge types that resources contain.
Look at levenshtein similarity between the resources students use and the knowledge type categories they use.
Visualize, anonymize and link data between the EDX server and the AppInventor Server.
Visualize resource usage for students over the time that the course is offered.
Visualize how students move between resources.
Determine resource reviews based on two metrics.
Visualize aggregated SCM and BKY usage.
Visualize nodes used by students in BKY and SCM.
Visualize aggregated SCM-BKY transitions.
Visualize aggregated signups and dropouts of students.
Generate necessary dataframes for other notebooks. Generates a serialized pickle file that the other notebooks load. Must run before other notebooks.
You must modify the variables in config.ini
, run user_and_course_dfs
. To get the necessary resource data from the course, you must run https://github.com/18goldr/web-crawler.
By default, all files will display the graphs inline. To save the graphs to a file, you must modify the variable to_file
in config.ini
to be True
. You may also modify to_file
anywhere within the scripts. This allows to save to file on a per-graph basis and is easier than having to reload external modules in Jupyter/IPython.
@inproceedings{gold2020analyzing,
title={Analyzing K-12 Blended MOOC Learning Behaviors},
author={Gold, Robert and Hemberg, Erik and O'Reilly, Una-May},
booktitle={Proceedings of the Seventh ACM Conference on Learning@ Scale},
pages={345--348},
year={2020}
}