10. Derive features from other features

Derive feature from other features¶

Features can also be created based on other features.

In this tutorial we will create a feature that measures the Z-Score of the Latest invoice Amount for the customer compared to that customer's invoices over a 28d period. This feature is derived from 3 features we created previously: CUSTOMER_Latest_invoice_Amount, CUSTOMER_Avg_of_invoice_Amount_28d and CUSTOMER_Std_of_invoice_Amount_28d.

In [1]:

import featurebyte as fb

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

catalog_name = "Grocery Dataset Tutorial"
catalog = fb.Catalog.activate(catalog_name)

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

15:31:33 | INFO     | SDK version: 1.0.2.dev46
15:31:33 | INFO     | No catalog activated.
15:31:33 | INFO     | Using profile: tutorial
15:31:33 | INFO     | Using configuration file at: /Users/gxav/.featurebyte/config.yaml
15:31:33 | INFO     | Active profile: tutorial (https://tutorials.featurebyte.com/api/v1)
15:31:33 | INFO     | SDK version: 1.0.2.dev46
15:31:33 | INFO     | No catalog activated.
15:31:33 | INFO     | Catalog activated: Grocery Dataset Tutorial

Get features we previously created and saved¶

In [2]:

customer_latest_invoice_amount = catalog.get_feature("CUSTOMER_Latest_invoice_Amount")
customer_avg_of_invoice_amount_28d = catalog.get_feature("CUSTOMER_Avg_of_invoice_Amount_28d")
customer_std_of_invoice_amount_28d = catalog.get_feature("CUSTOMER_Std_of_invoice_Amount_28d")

customer_latest_invoice_amount = catalog.get_feature("CUSTOMER_Latest_invoice_Amount") customer_avg_of_invoice_amount_28d = catalog.get_feature("CUSTOMER_Avg_of_invoice_Amount_28d") customer_std_of_invoice_amount_28d = catalog.get_feature("CUSTOMER_Std_of_invoice_Amount_28d")

Derive a z-score feature from saved features¶

In [3]:

customer_latest_invoice_amount_Z_score_to_invoice_amount_28d = (
    customer_latest_invoice_amount
    - customer_avg_of_invoice_amount_28d
) / customer_std_of_invoice_amount_28d
# Give a name to new feature
customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.name = \
"CUSTOMER_Latest_invoice_Amount_Z_Score_to_invoice_Amount_28d"

customer_latest_invoice_amount_Z_score_to_invoice_amount_28d = ( customer_latest_invoice_amount - customer_avg_of_invoice_amount_28d ) / customer_std_of_invoice_amount_28d # Give a name to new feature customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.name = \ "CUSTOMER_Latest_invoice_Amount_Z_Score_to_invoice_Amount_28d"

Preview feature¶

In [4]:

customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.primary_entity

customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.primary_entity

Out[4]:

[<featurebyte.api.entity.Entity at 0x13eda6b40>
 {
   'name': 'customer',
   'created_at': '2024-04-26T07:28:15.128000',
   'updated_at': '2024-04-26T07:28:24.021000',
   'description': None,
   'serving_names': [
     'GROCERYCUSTOMERGUID'
   ],
   'catalog_name': 'Grocery Dataset Tutorial'
 }]

In [5]:

# Get observation table: 'Preview Table with 10 items'
preview_table = catalog.get_observation_table("Preview Table with 10 items")

# Get observation table: 'Preview Table with 10 items' preview_table = catalog.get_observation_table("Preview Table with 10 items")

In [6]:

# Preview CUSTOMER_Age
customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.preview(preview_table)

# Preview CUSTOMER_Age customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.preview(preview_table)

Out[6]:

	POINT_IN_TIME	GROCERYINVOICEITEMGUID	CUSTOMER_Latest_invoice_Amount_Z_Score_to_invoice_Amount_28d
0	2022-11-03 14:39:00	9faf5936-d4bb-4709-a530-c1624ec003a5	2.039603
1	2022-10-05 14:30:03	fc4769cf-41d6-4fb0-9cfc-fe30502cfa18	-0.153076
2	2022-08-16 15:37:21	6781acc6-652d-4867-a138-2d8adb278886	-0.849703
3	2023-03-13 21:11:43	cfc620a2-0054-4b4b-99d5-040cf87cfe2d	-0.799217
4	2023-03-19 15:22:18	dcdb7c98-dd62-4287-b432-bfe3a2317ebc	0.958202
5	2022-11-18 10:07:49	ae2ccf38-4e5d-4c76-b1e5-04f12307e45b	-0.586120
6	2022-11-09 18:07:20	37122b82-478b-4d9e-b236-69629b592c0b	-1.000000
7	2023-02-15 21:10:22	5daf8edb-8625-4653-aaa0-9ac03df92017	-1.168187
8	2023-02-27 13:27:47	b807e05c-ff1c-4fb3-a760-e0e8ce29c859	-0.512459
9	2023-04-19 15:25:05	8a60f455-aff1-4b4e-8c63-9ab89df2715d	-1.327036

Save feature¶

In [7]:

customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.save()

customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.save()

Done! |████████████████████████████████████████| 100% in 6.1s (0.16%/s)         

Add description and view definition file¶

In [8]:

# Add description
customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.update_description(
	"Z-Score of the Latest invoice Amount for the customer compared to customer invoices over a 28d period"
)
# See feature definition file
customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.definition

# Add description customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.update_description( "Z-Score of the Latest invoice Amount for the customer compared to customer invoices over a 28d period" ) # See feature definition file customer_latest_invoice_amount_Z_score_to_invoice_amount_28d.definition

Out[8]:

# Generated by SDK version: 1.0.2.dev46
from bson import ObjectId
from featurebyte import ColumnCleaningOperation
from featurebyte import DisguisedValueImputation
from featurebyte import EventTable
from featurebyte import FeatureJobSetting
from featurebyte import ValueBeyondEndpointImputation


# event_table name: "GROCERYINVOICE"
event_table = EventTable.get_by_id(ObjectId("662b5775aa13c89fa14554e1"))
event_view = event_table.get_view(
    view_mode="manual",
    drop_column_names=["record_available_at"],
    column_cleaning_operations=[
        ColumnCleaningOperation(
            column_name="Amount",
            cleaning_operations=[
                DisguisedValueImputation(
                    imputed_value=None, disguised_values=[-99.0, -98.0]
                ),
                ValueBeyondEndpointImputation(
                    type="less_than", end_point=0.0, imputed_value=0.0
                ),
                ValueBeyondEndpointImputation(
                    type="greater_than", end_point=2000.0, imputed_value=2000.0
                ),
            ],
        )
    ],
)
grouped = event_view.groupby(
    by_keys=["GroceryCustomerGuid"], category=None
).aggregate_over(
    value_column="Amount",
    method="latest",
    windows=[None],
    feature_names=["CUSTOMER_Latest_invoice_Amount"],
    feature_job_setting=FeatureJobSetting(
        blind_spot="120s", frequency="3600s", time_modulo_frequency="120s"
    ),
    skip_fill_na=True,
)
feat = grouped["CUSTOMER_Latest_invoice_Amount"]
grouped_1 = event_view.groupby(
    by_keys=["GroceryCustomerGuid"], category=None
).aggregate_over(
    value_column="Amount",
    method="avg",
    windows=["28d"],
    feature_names=["CUSTOMER_Avg_of_invoice_Amount_28d"],
    feature_job_setting=FeatureJobSetting(
        blind_spot="120s", frequency="3600s", time_modulo_frequency="120s"
    ),
    skip_fill_na=True,
)
feat_1 = grouped_1["CUSTOMER_Avg_of_invoice_Amount_28d"]
grouped_2 = event_view.groupby(
    by_keys=["GroceryCustomerGuid"], category=None
).aggregate_over(
    value_column="Amount",
    method="std",
    windows=["28d"],
    feature_names=["CUSTOMER_Std_of_invoice_Amount_28d"],
    feature_job_setting=FeatureJobSetting(
        blind_spot="120s", frequency="3600s", time_modulo_frequency="120s"
    ),
    skip_fill_na=True,
)
feat_2 = grouped_2["CUSTOMER_Std_of_invoice_Amount_28d"]
feat_3 = (feat - feat_1) / feat_2
feat_3.name = "CUSTOMER_Latest_invoice_Amount_Z_Score_to_invoice_Amount_28d"
output = feat_3
output.save(_id=ObjectId("662b5858869e92ad5091c274"))

Concepts in this tutorial¶

• Feature Transforms

SDK reference for¶

• Catalog.get_feature()