Skip to content

Drivers (Brand Analysis)

Drivers Core

adele.modeling.drivers.core.DriversS3

adele.modeling.drivers.core.DriversS3Data (DriversS3)

Drivers Single

adele.modeling.drivers.single.core.DriversSingle (DriversS3Data)

Base class for a single driver estimation

Examples:

>>> from adele.modeling.drivers.single.core import DriversSingle

>>> ds = DriversSingle("dev")
>>> setattr(ds, "drivers_id", DriverId)
>>> setattr(ds, "job_id", JobId)

>>> ds.read_impact()
read_impact(self)

Read impact scores from the drivers run

adele.modeling.drivers.single.estimation.core.DriversLoad (DriversSingle)

Load the payload information and assign the attributes

data_stacked_filtered_T2B property readonly

data_stacked filtered as specified by the model_payload's filter and only T2B columns

adele.modeling.drivers.single.estimation.aggregates.DriversAggregates (DriversLoad)

Calculate aggregate statistics off the stacked data specific to the drivers run

Examples:

>>> from adele.modeling.drivers.job.payload_generator import DriversJobPayloadGenerator
>>> from adele.modeling.drivers.single.estimation.aggregates import DriversAggregates
>>> gen = DriversJobPayloadGenerator(JobId,"dev")

>>> DriverId = list(gen.all_payloads.keys())[10]
>>> _payload = gen.all_payloads[DriverId]

>>> aggregates = DriversAggregates("dev")
>>> setattr(aggregates, "payload", _payload)
>>> aggregates.retrieve_tables(["correlation_table",
                                    "performance_table",
                                    "brand_base_table",
                                    "double_centered_table",
                                    "subgroup_base_table"])
correlation_table property readonly

Correlation Table

performance_table property readonly

Performance Table

brand_base_table property readonly

Base Table

double_centered_table property readonly

Double centered table

subgroup_base_table property readonly

Subgroup base

read_aggregates(self, which_tables)

Specify which tables to retrieve

adele.modeling.drivers.single.estimation.impacts.DriversImpact (DriversLoad)

Driver class for a single model

Examples:

>>> from adele.processing.stacking.process.core import StackerProcess
>>> from adele.modeling.drivers.job.payload_generator import DriversJobPayloadGenerator
>>> from adele.modeling.drivers.single.estimation.impacts import DriversImpact

>>> m = StackerProcess('dev','enctest', username, password)
>>> setattr(m, "payload", spec)
>>> m.stack()

>>> gen = DriversJobPayloadGenerator(JobId,"dev")

>>> for DriverId in list(gen.all_payloads.keys()):

print(DriverId)

_payload = gen.all_payloads[DriverId]

driver = DriversImpact(_payload, "dev")
driver.execute_impact()
execute_impact(self)

Shortcut to get the right model to run and save results to S3

Drivers Job

adele.modeling.drivers.job.core.DriversJob (DriversS3Data)

Create payloads for individual models based off stacked data/dictionary info

Examples:

>>> from adele.modeling.drivers.job.core import DriversJob
>>> JobId = '2805ff44-c710-4212-97de-4c6eed8db37c'
>>> job = DriversJob(JobId,"dev")
>>> job.poll_job()

adele.modeling.drivers.job.payload_generator.DriversJobPayloadGenerator (DriversJob)

Create payloads for individual models based off stacked data/dictionary info

Examples:

>>> from adele.modeling.drivers.job.payload_generator import DriversJobPayloadGenerator
>>> JobId = '2805ff44-c710-4212-97de-4c6eed8db37c'
>>> gen = DriversJobPayloadGenerator(JobId,"dev")
>>> gen.all_payloads
read_payloads(self)

Read payloads from S3 if they've been run already

generate_payloads(self)

Wrapper for the internal methods

create_combos(self, fix_t2b=True)

Create all combinations of IV/DV recodes, IVs, Dvs, Subgroups, Filters, Model Types and Estimators

Parameters:

Name Type Description Default
fix_t2b bool

this will fix the performance and double centered table to top 2 box for all drivers runs

 True
prepare_payloads_from_combos(self)

Convert combos from dataframe to dictionary row-wise

Last update: 2022-06-07
Back to top