fusetools.analytics_tools.Pandas

class fusetools.analytics_tools.Pandas

Bases: object

Functions for running analytical Pandas operations

Methods

append_window_agg	Joins a window function’s aggregation to a Pandas DataFrame.
find_na_holder	Returns the combination of 2 de-duped columns in a Pandas DataFrame.
period_comp	Creates a snapshot comparison between two periods.
period_start_dt	Returns the first day of a year or month for a Pandas Series.
ptd_measure	Creates a ‘Period to date’ aggregation.
yoy_comp	Computes a YoY cumulative YTD comparison across for a given week.

classmethod append_window_agg(df, dim, metric, metric_agg, comp_col=False)

Joins a window function’s aggregation to a Pandas DataFrame.

Parameters

• df – Pandas DataFrame.

• dim – Column for which to partition of data by.

• metric – Column to aggregate.

• metric_agg – Type of calculation to perform.

• comp_col – Flag of whether or not to create a comparison column.

Returns

Pandas DataFrame with a window function’s aggregation.

classmethod find_na_holder(df, col, col_new)

Returns the combination of 2 de-duped columns in a Pandas DataFrame.

Parameters

• df – Pandas DataFrame

• col – Original column.

• col_new – New column.

Returns

Combination of 2 de-duped columns in a Pandas DataFrame.

classmethod period_comp(df, period_field, val_fields, dim=False, val_field_suffix=False, hist=False)

Creates a snapshot comparison between two periods.

Parameters

• df – Pandas DataFrame.

• period_field – Column with period to compare across

• val_fields – List of columns with numeric values to compare

• dim – Column with dimension to group across (optional)

• val_field_suffix – Suffix for value field to add to final dataset (optional)

• hist – Include history flag, returns all periods if True, otherwise just the most recent two periods

Returns

Comparison Pandas DataFrame

classmethod period_start_dt(df)

Returns the first day of a year or month for a Pandas Series.

Parameters

df – Pandas DataFrame.

Returns

First day of year or month for Pandas Series.

classmethod ptd_measure(df, period, val_fields, kpi, dim=False)

Creates a ‘Period to date’ aggregation.

Parameters

• df – Pandas DataFrame

• period – Type of period (year, month)

• val_fields – Columns to aggregate.

• kpi – Type of aggregation to perform.

• dim – Dimension to group comparison by (Optional).

Returns

Pandas DataFrame with PTD measure.

classmethod yoy_comp(df, val_dict, dim=False, hist=False)

Computes a YoY cumulative YTD comparison across for a given week.

Parameters

• df – Pandas DataFrame.

• val_dict – Column and aggregation type specification.

• dim – Dimension to group comparison by (Option).

• hist – Flag of whether to keep all historical date combinations.

Returns

Pandas DataFrame with YoY cumulative YTD comparison.