FILE: docs/api/pipelines.md¶
Pipelines API Reference¶
High‑level workflows that combine multiple DataKit functions into single calls.
quick_eda(df, show_plots=True, save_dir=None, export_profile=False, verbose=True)¶
Run a full EDA pipeline: profile, missing recommendations, distributions, correlation.
Parameters:
- df – pandas DataFrame
- show_plots – display plots if True
- save_dir – directory to save plots (None = don't save)
- export_profile – export HTML profile report (saves as profile.html)
- verbose – print reports from sub‑functions
Returns: Dict with keys:
- 'profile' – the profile dict from profile_df()
- 'corr_matrix' – correlation matrix (or None if <2 numeric columns)
Example:
result = dk.quick_eda(df)
result = dk.quick_eda(df, save_dir='eda_output/', export_profile=True)
groupby_chart(df, by, metric, agg='mean', chart_type='bar', annotate=True, title=None, figsize=(10, 5), show=True, save=None, dpi=150)¶
Group by a column, aggregate a metric, and plot the result.
Parameters:
- df – pandas DataFrame
- by – column name or list of columns to group by
- metric – column name to aggregate
- agg – aggregation function ('mean', 'sum', 'count', 'median', or a callable)
- chart_type – 'bar' (vertical) or 'barh' (horizontal)
- annotate – add value labels on bars
- title – chart title (auto‑generated if None)
- figsize – figure size (width, height)
- show – call plt.show()
- save – file path to save figure (None = don't save)
- dpi – resolution for saved figure
Returns: matplotlib Axes object
Example:
dk.groupby_chart(df, by='region', metric='sales', agg='sum')
dk.groupby_chart(df, by=['year', 'quarter'], metric='revenue', agg='mean', chart_type='barh')