Visualize the MOOC learner features with MOOC-Learner-Visualized (MLV)

(see MOOC-Learner-Docker/visualized_base_img )

See MOOC-Learner-Docker

Entry point is autorun.py. Configuration is done with config/*yml, see e.g. config/sample_config.yml.

As a backend, run MOOC-Learner-Curated only once, run MOOC-Learner-Quantified only when necessary by command line then run MOOC-Learner-Visualized (the Flask server)

MOOC-Learner-Visualized serves as an interface to visualize features populated in the moocdb. It contains three parts:

• Fetching: which fetches feature columns from feature tables
• Processing: provides several processing functions on data columns, such as filtering, mapping, and statistics calculation. Finally, it performs an inner join on relevant feature rows to form a data frame for rendering
• Rendering: provides an interface to configure and render interactive and static plots. The template engine can automatically generate html input forms and uses javascript to collect the user input. Adding a new type of plot is as easy as adding a new dictionary describing the configuration form and write up a new drawing function, which takes a pandas dataframe and the all form input as parameters and return a bokeh figure instance. Currently we have scatter plot, histogram as templates for interactive plots and scatter matrix as the template for static plots. For interactive plots, the user can select or filter a specific feature column with a slider.