Data logo human learn: data

1 2 3 4 5 6 7 8

Let's make some drawings without supplying a label column.

from hulearn.experimental.interactive import InteractiveCharts

# Note that we now supply a list of labels to draw.
charts = InteractiveCharts(df, labels=["group_one", "group_two"])
charts.add_chart(x="bill_length_mm", y="bill_depth_mm")

These drawn features can also be applied as a featurization step. Both scikit-learn pipelines and pandas pipelines are supported.

from hulearn.preprocessing import InteractivePreprocessor
tfm = InteractivePreprocessor(json_desc=charts.data())

# The flow for scikit-learn
tfm.fit(df).transform(df)

# The flow for pandas
df.pipe(tfm.pandas_pipe)