6. Formulate Use Case

Why Create a Use Case?¶

Creating a Use Case in FeatureByte, while optional, offers significant benefits. It not only defines the goal and application of your prediction model but also helps organize your training data and improve post-deployment feature monitoring.

Plus, for FeatureByte Enterprise users, it's used by FeatureByte's Feature Ideation engine to suggest features tailored to your use cases.

We'll create one Use Case for our Loan Applications Dataset:

Loan Default by client: Predict clients' payment difficulties the next 6 months for new loan based on its early stage application data and prior credit history.

We will define a Context, establish a Target and then link the Use Case with the Context and Target.

In [1]:

import featurebyte as fb

# Set your profile to the tutorial environment
fb.use_profile("tutorial")

catalog_name = "Loan Applications Dataset SDK Tutorial"
catalog = fb.Catalog.activate(catalog_name)

import featurebyte as fb # Set your profile to the tutorial environment fb.use_profile("tutorial") catalog_name = "Loan Applications Dataset SDK Tutorial" catalog = fb.Catalog.activate(catalog_name)

14:08:12 | INFO     | SDK version: 3.0.1.dev45
INFO    :featurebyte:SDK version: 3.0.1.dev45
14:08:12 | INFO     | No catalog activated.
INFO    :featurebyte:No catalog activated.
14:08:12 | INFO     | Using profile: tutorial
INFO    :featurebyte:Using profile: tutorial
14:08:12 | INFO     | Using configuration file at: /Users/gxav/.featurebyte/config.yaml
INFO    :featurebyte:Using configuration file at: /Users/gxav/.featurebyte/config.yaml
14:08:12 | INFO     | Active profile: tutorial (https://tutorials.featurebyte.com/api/v1)
INFO    :featurebyte:Active profile: tutorial (https://tutorials.featurebyte.com/api/v1)
14:08:13 | INFO     | SDK version: 3.0.1.dev45
INFO    :featurebyte:SDK version: 3.0.1.dev45
14:08:13 | INFO     | No catalog activated.
INFO    :featurebyte:No catalog activated.
14:08:13 | INFO     | Catalog activated: Loan Applications Dataset SDK Tutorial
INFO    :featurebyte.api.catalog:Catalog activated: Loan Applications Dataset SDK Tutorial
16:06:14 | INFO     | No catalog activated.
16:06:14 | INFO     | Catalog activated: Grocery Dataset Tutorial

Create a Context¶

A Context defines the scope and circumstances in which features are expected to be served.

Let's describe the Context for our use case:

• Context Name: New Loan Application
• Description: Loan application under review.
• Primary Entity: New Application

In [2]:

context_name = "New Loan Application"
fb.Context.create(
    context_name,
    primary_entity=["New Application"],
    description="Loan application under review.",
)

context_name = "New Loan Application" fb.Context.create( context_name, primary_entity=["New Application"], description="Loan application under review.", )

Out[2]:

Context

name	New Loan Application
created_at	2025-06-02 05:50:13
updated_at	None
description	Loan application under review.
author	None
primary_entities	name serving_names catalog_name 0 New Application [SK_ID_CURR] Loan Applications Dataset SDK Tutorial
default_eda_table	None
default_preview_table	None
associated_use_cases	[]

Creating a Target¶

For this use case, we'll use the descriptive approach, as the target cannot be derived directly from the data and must be provided alongside the observation tables.

For an example of a target computed by the SDK, check out the Grocery SDK Tutorial.

In [3]:

target_name = "Loan_Default"
target = fb.TargetNamespace.create(
    name=target_name,
    window="182d",
    dtype="INT",
    primary_entity=["New Application"],
    target_type=fb.TargetType.CLASSIFICATION,
)

target_name = "Loan_Default" target = fb.TargetNamespace.create( name=target_name, window="182d", dtype="INT", primary_entity=["New Application"], target_type=fb.TargetType.CLASSIFICATION, )

Create a Use Case¶

Let's link our Use Case with the Context and Target.

In [4]:

use_case_name = "Loan Default by client"
use_case_description = (
    "Predict clients' payment difficulties the next 6 months for new loan based on its "
    "application data and prior credit history."
)
fb.UseCase.create(
    name=use_case_name,
    target_name=target_name,
    context_name=context_name,
    description=use_case_description,
)

use_case_name = "Loan Default by client" use_case_description = ( "Predict clients' payment difficulties the next 6 months for new loan based on its " "application data and prior credit history." ) fb.UseCase.create( name=use_case_name, target_name=target_name, context_name=context_name, description=use_case_description, )

Out[4]:

Use Case

name	Loan Default by client
created_at	2025-06-02 05:50:15
updated_at	None
description	Predict clients' payment difficulties the next 6 months for new loan based on its application data and prior credit history.
author	None
primary_entities	name serving_names catalog_name 0 New Application [SK_ID_CURR] Loan Applications Dataset SDK Tutorial
context_name	New Loan Application
target_name	Loan_Default
default_eda_table	None
default_preview_table	None

List targets, contexts and use cases in the catalog¶

In [5]:

catalog.list_targets()

catalog.list_targets()

Out[5]:

	id	name	dtype	entities	created_at
0	None	Loan_Default	INT	[New Application]	2025-06-02T05:50:14.874000

In [6]:

catalog.list_contexts()

catalog.list_contexts()

Out[6]:

	id	name	primary_entity_ids	description
0	683d3b95e1689f620853fa3d	New Loan Application	[683d3ad661061d4da44450d0]	Loan application under review.

In [7]:

catalog.list_use_cases()

catalog.list_use_cases()

Out[7]:

	id	name	default_preview_table_name	default_eda_table_name	description
0	683d3b97e1689f620853fa40	Loan Default by client	None	None	Predict clients' payment difficulties the next...

Concepts discussed in this tutorial¶

• More about feature definition file

SDK reference for¶

• Target
• Context
• Use Case
• View.groupby()
• GroupBy.forward_aggregate()
• Aggregation functions
• Target.definition