Visualization Utilities

Use the utilities below to visualize the comparison between real and synthetic data. You can access these from the sdmetrics.visualization module.

Tip! All visualizations are interactive. If you're using an iPython notebook, you can zoom, pan, toggle legends and take screenshots.

Compare a synthetic column & real column (1D)

get_column_plot

Use this utility to visualize a real column against the same synthetic column. You can plot any column of type: boolean, categorical, datetime or numerical.

• (required) real_data: A pandas.DataFrame containing the table of your real data. To skip plotting the real data, input None.

• (required) synthetic_data: A pandas.DataFrame containing the synthetic data. To skip plotting the synthetic data, input None.

• (required) column_name: The name of the column you want to plot.

• plot_type: The type of plot to create

  • (default) None: Determine the type of plot to create based on the data.

  • 'distplot': Plot the data as a smooth, continuous distribution. Use this for continuous columns.

  • 'bar': Plot the data as discrete bars. Use this for discrete columns.

Returns: A plotly.Figure object

from sdmetrics.visualization import get_column_plot

fig = get_column_plot(
    real_data=real_table,
    synthetic_data=synthetic_table,
    column_name='high_perc',
    plot_type='distplot'
)

fig.show()

Compare a pair of synthetic columns & real columns (2D)

utils.get_column_pair_plot

Use this utility to visualize the trends between a pair of columns for real and synthetic data. You can plot any 2 columns of type: boolean, categorical, datetime or numerical. The columns do not have to the be the same type.

• (required) real_data: A pandas.DataFrame containing the table of your real data. To skip plotting the real data, input None.

• (required) synthetic_data: A pandas.DataFrame containing the synthetic data. To skip plotting the synthetic data, input None.

• (required) column_names: A list containing the names of the 2 columns you want to plot.

• plot_type: The type of plot to create

  • (default) None: Determine the type of plot to create based on the data.

  • 'scatter': Plot each data point in 2D space as a scatter plot. Use this to compare a pair of continuous columns.

  • 'box': Plot the data as one or more box plot. Use this to compare a continuous column with a discrete column.

  • 'heatmap': Plot a side-by-side headmap of the data's categories. Use this to compare a pair of discrete columns.

Returns: A plotly.Figure object

from sdmetrics.visualization import get_column_pair_plot

fig = get_column_pair_plot(
    real_data=real_table,
    synthetic_data=synthetic_table,
    column_names=['mba_perc', 'degree_perc'],
    plot_type='scatter'
    
)

fig.show()

Various types of plots are possible based on the types of data you provide

Visualize the cardinality of a relationship

utils.get_cardinality_plot

Use this utility to visualize the cardinality of parent-child relationship. The cardinality is the # of children that each parent row has. Your cardinality may be fixed (eg. every parent has exactly 2 children) or variable (eg. every parent has 1-3 children).

• (required) real_data: A dictionary mapping the name of each table to a pandas.DataFrame containing the real data for that table. To skip plotting the real data, input None.

• (required) synthetic_data: A dictionary mapping the name of each table to a pandas.DataFrame containing the synthetic data for that table. To skip plotting the synthetic data, input None.

• (required) parent_table_name: The string name of the parent table in the relationship

• (required) child_table_name: The string name of the child table in the relationship

• (required) parent_primary_key: The string name of the parent table's primary key

• (required) child_foreign_key: The string name of the column in the child table that refers to the parent's primary key

• plot_type: The type of plot to create

  • (default) None: Determine the type of plot to create based on the data.

  • 'distplot': Plot the data as a smooth, continuous distribution

  • 'bar': Plot the data as discrete bars

Returns: A plotly.Figure object

from sdmetrics.visualization import get_cardinality_plot

fig = get_cardinality_plot(
    real_data=real_tables,
    synthetic_data=synthetic_tables,
    parent_table_name='users',
    child_table_name='sessions',
    parent_primary_key='user_id'
    child_foreign_key='user_id',
    plot_type='bar'
)

fig.show()