DotPlotter Example

This example shows how to use the DotPlotter class to generate a bar chart to study the estimated mean of the learner’s knowledge and our confidence level (via error bars).

It is similar to the previous example of the BarPlotter. The difference is that the bars have been replaced with dots.

In this example, we use the NoveltyClassifier to build a representation of the learner’s knowledge. You could also use other classifiers like KnowledgeClassifier (for building knowledge representation) or InterestClassifier (for building interest representation).

from truelearn import learning, datasets, models
from truelearn.utils import visualisations

import plotly.io as pio

# You can also use a custom knowledge component
# if it follows the protocol of history aware knowledge component
data, _, _ = datasets.load_peek_dataset(
    test_limit=0, kc_init_func=models.HistoryAwareKnowledgeComponent, verbose=False
)

# select a learner from data
_, learning_events = data[12]

classifier = learning.NoveltyClassifier()
for event, label in learning_events:
    classifier.fit(event, label)

plotter = visualisations.DotPlotter()

# you can control whether to include history data
# in the plot. If you use `history=True`, when you hover
# your mouse over the dot, you can see statistics about
# the total number of videos watched and the time the learner watched the last video
plotter.plot(classifier.get_learner_model().knowledge, top_n=10, history=True)

# you can also use plotter.show()
# which is a shorthand for calling pio
pio.show(plotter.figure)