API Reference
Public classes, methods, and attributes documented by the current lochan-eda API.
AutomatedEDA
High-level interface for the complete preprocessing workflow.
from lochan_eda import AutomatedEDA
eda = AutomatedEDA()Method / attributeDescription
prepare()High-level interface for the common tabular preprocessing workflow.fit()Fit the preprocessing workflow using training data.transform()Apply the fitted preprocessing workflow to data.Profiler
Dataset-level analysis interface.
from lochan_eda import Profiler
profile = Profiler(df)Method / attributeDescription
overview()Dataset-level inspection.numericalAccess to the numerical analysis component.categoricalAccess to the categorical analysis component.Numerical
Numerical feature analysis and preprocessing.
from lochan_eda import Numerical
numerical = Numerical(df)Method / attributeDescription
fit()Learn data behaviour.transform()Execute Learnt things.summary()Statistical summaries.plot()Visualization.Categorical
Categorical feature analysis and preprocessing.
from lochan_eda import Categorical
categorical = Categorical(df)Method / attributeDescription
fit()Learn data behaviour.transform()Execute Learnt things.summary()Statistical summaries.plot()Visualization.Missing
Missing-value visualization.
Method / attributeDescription
plot()Missing-value visualization.Report
Generate a shareable analysis report.
Method / attributeDescription
save()Save the analysis report.High-level API
Use AutomatedEDA when the goal is to move efficiently from a DataFrame to model-ready data.
eda = AutomatedEDA()
X_train, X_test, y_train, y_test = eda.prepare(
df,
target="target"
)Component-level API
Use Profiler, Numerical, and Categorical when explicit control or inspection is required.
profile = Profiler(df)
profile.overview()
profile.numerical
profile.categoricalPackage structure
lochan_eda/
│
├── automated_eda/
├── profiler/
├── numerical/
├── categorical/
├── missing/
└── report/Dependencies
pandasNumPyscikit-learnmatplotlibreportlab
Contributing
Keep the public API consistent with the existing design. Add or update tests for behavioural changes. Keep preprocessing decisions explicit and reproducible. Avoid unnecessary dependencies. Document new public functionality.