9. Create Window Aggregate Features

Create window aggregate features¶

Next feature type we will consider is window aggregate feature. These are features generated by aggregating data within specific time frame.

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)

16:07:43 | WARNING  | Service endpoint is inaccessible: http://featurebyte-server:8088
16:07:43 | INFO     | Using profile: tutorial
16:07:43 | INFO     | Using configuration file at: /Users/gxav/.featurebyte/config.yaml
16:07:43 | INFO     | Active profile: tutorial (https://tutorials.featurebyte.com/api/v1)
16:07:43 | WARNING  | Remote SDK version (1.1.0.dev7) is different from local (1.1.0.dev1). Update local SDK to avoid unexpected behavior.
16:07:43 | INFO     | No catalog activated.
16:07:43 | INFO     | Catalog activated: Grocery Dataset Tutorial

In [2]:

# Set desired windows
windows = ['14d', '28d']

# Set desired windows windows = ['14d', '28d']

Do window aggregation from INVOICEITEMS¶

Let's start with some aggregations from the items view and create features for the interaction between Customer and Product Group.

In [3]:

# Get view from GROCERYPRODUCT dimension table.
groceryproduct_view = catalog.get_view("GROCERYPRODUCT")
# Get view from INVOICEITEMS item table.
invoiceitems_view = catalog.get_view("INVOICEITEMS")
# Join GROCERYPRODUCT view to INVOICEITEMS view.
invoiceitems_view = invoiceitems_view.join(groceryproduct_view, rsuffix="")

# Get view from GROCERYPRODUCT dimension table. groceryproduct_view = catalog.get_view("GROCERYPRODUCT") # Get view from INVOICEITEMS item table. invoiceitems_view = catalog.get_view("INVOICEITEMS") # Join GROCERYPRODUCT view to INVOICEITEMS view. invoiceitems_view = invoiceitems_view.join(groceryproduct_view, rsuffix="")

In [4]:

# Group GROCERYINVOICE view by customer entity (GroceryCustomerGuid) and productgroup entity (ProductGroup).
invoiceitems_view_by_customer_x_productgroup = invoiceitems_view.groupby(
    ['GroceryCustomerGuid', 'ProductGroup']
)

# Group GROCERYINVOICE view by customer entity (GroceryCustomerGuid) and productgroup entity (ProductGroup). invoiceitems_view_by_customer_x_productgroup = invoiceitems_view.groupby( ['GroceryCustomerGuid', 'ProductGroup'] )

In [5]:

# Get Sum of TotalCost for the customer x productgroup over time.
customer_productgroup_sum_of_totalcost_14d_28d = \
invoiceitems_view_by_customer_x_productgroup.aggregate_over(
    "TotalCost", method="sum",
    feature_names=[
        "CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost"
        + "_" + w for w in windows
    ],
    windows=windows
)

# Get Sum of TotalCost for the customer x productgroup over time. customer_productgroup_sum_of_totalcost_14d_28d = \ invoiceitems_view_by_customer_x_productgroup.aggregate_over( "TotalCost", method="sum", feature_names=[ "CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost" + "_" + w for w in windows ], windows=windows )

In [6]:

# Get Latest Interaction between Customer and ProductGroup
customer_x_productgroup_latest_timestamp = \
invoiceitems_view_by_customer_x_productgroup.aggregate_over(
    "Timestamp", method="latest",
    feature_names=["CUSTOMER_x_PRODUCTGROUP_Latest_Timestamp"],
    windows=[None]
)["CUSTOMER_x_PRODUCTGROUP_Latest_Timestamp"]

# Get Latest Interaction between Customer and ProductGroup customer_x_productgroup_latest_timestamp = \ invoiceitems_view_by_customer_x_productgroup.aggregate_over( "Timestamp", method="latest", feature_names=["CUSTOMER_x_PRODUCTGROUP_Latest_Timestamp"], windows=[None] )["CUSTOMER_x_PRODUCTGROUP_Latest_Timestamp"]

In [7]:

# Create recency feature: Time Since Latest Interaction between Customer and Product Group
customer_x_productgroup_time_since_latest_timestamp = (
    fb.RequestColumn.point_in_time()
    - customer_x_productgroup_latest_timestamp
).dt.hour
customer_x_productgroup_time_since_latest_timestamp.name = \
"CUSTOMER_x_PRODUCTGROUP_Time_Since_Latest_Timestamp"

# Create recency feature: Time Since Latest Interaction between Customer and Product Group customer_x_productgroup_time_since_latest_timestamp = ( fb.RequestColumn.point_in_time() - customer_x_productgroup_latest_timestamp ).dt.hour customer_x_productgroup_time_since_latest_timestamp.name = \ "CUSTOMER_x_PRODUCTGROUP_Time_Since_Latest_Timestamp"

Do window aggregation from GROCERYINVOICE¶

Now, let's do some aggregations on the invoices view for the Customer entity.

In [8]:

# Get view from GROCERYINVOICE event table.
groceryinvoice_view = catalog.get_view("GROCERYINVOICE")

# Get view from GROCERYINVOICE event table. groceryinvoice_view = catalog.get_view("GROCERYINVOICE")

In [9]:

# Group GROCERYINVOICE view by customer entity (GroceryCustomerGuid).
groceryinvoice_view_by_customer = groceryinvoice_view.groupby(['GroceryCustomerGuid'])

# Group GROCERYINVOICE view by customer entity (GroceryCustomerGuid). groceryinvoice_view_by_customer = groceryinvoice_view.groupby(['GroceryCustomerGuid'])

In [10]:

# Get Latest invoice Amount for the customer
customer_latest_invoice_amount = groceryinvoice_view_by_customer.aggregate_over(
    "Amount", method="latest",
    feature_names=["CUSTOMER_Latest_invoice_Amount"],
    windows=[None]
)["CUSTOMER_Latest_invoice_Amount"]

# Get Latest invoice Amount for the customer customer_latest_invoice_amount = groceryinvoice_view_by_customer.aggregate_over( "Amount", method="latest", feature_names=["CUSTOMER_Latest_invoice_Amount"], windows=[None] )["CUSTOMER_Latest_invoice_Amount"]

In [11]:

# Get Count of invoices for the customer
customer_count_of_invoice_14d_28d = groceryinvoice_view_by_customer.aggregate_over(
    method="count",
    feature_names=[
        "CUSTOMER_Count_of_invoice"
        + "_" + w for w in windows
    ],
    windows=windows
)

# Get Count of invoices for the customer customer_count_of_invoice_14d_28d = groceryinvoice_view_by_customer.aggregate_over( method="count", feature_names=[ "CUSTOMER_Count_of_invoice" + "_" + w for w in windows ], windows=windows )

In [12]:

# Get Avg of Amount for the customer over time.
customer_avg_of_invoice_amount_14d_28d = groceryinvoice_view_by_customer.aggregate_over(
    "Amount", method="avg",
    feature_names=[
        "CUSTOMER_Avg_of_invoice_Amount"
        + "_" + w for w in windows
    ],
    windows=windows
)

# Get Avg of Amount for the customer over time. customer_avg_of_invoice_amount_14d_28d = groceryinvoice_view_by_customer.aggregate_over( "Amount", method="avg", feature_names=[ "CUSTOMER_Avg_of_invoice_Amount" + "_" + w for w in windows ], windows=windows )

In [13]:

# Get Std of Amount for the customer over time.
customer_std_of_invoice_amount_14d_28d = groceryinvoice_view_by_customer.aggregate_over(
    "Amount", method="std",
    feature_names=[
        "CUSTOMER_Std_of_invoice_Amount"
        + "_" + w for w in windows
    ],
    windows=windows
)

# Get Std of Amount for the customer over time. customer_std_of_invoice_amount_14d_28d = groceryinvoice_view_by_customer.aggregate_over( "Amount", method="std", feature_names=[ "CUSTOMER_Std_of_invoice_Amount" + "_" + w for w in windows ], windows=windows )

Preview a feature group¶

For convenience, we can create a feature group to preview/save all features we just created.

In [14]:

feature_group = fb.FeatureGroup([
    customer_x_productgroup_time_since_latest_timestamp,
    customer_productgroup_sum_of_totalcost_14d_28d,
    customer_latest_invoice_amount,
    customer_count_of_invoice_14d_28d,
    customer_avg_of_invoice_amount_14d_28d,
    customer_std_of_invoice_amount_14d_28d,
])

feature_group = fb.FeatureGroup([ customer_x_productgroup_time_since_latest_timestamp, customer_productgroup_sum_of_totalcost_14d_28d, customer_latest_invoice_amount, customer_count_of_invoice_14d_28d, customer_avg_of_invoice_amount_14d_28d, customer_std_of_invoice_amount_14d_28d, ])

In [15]:

# Check the primary entity of the feature group. It should be the interaction Customer x ProductGroup.
feature_group.primary_entity

# Check the primary entity of the feature group. It should be the interaction Customer x ProductGroup. feature_group.primary_entity

Out[15]:

[<featurebyte.api.entity.Entity at 0x10687f080>
 {
   'name': 'customer',
   'created_at': '2024-06-12T08:05:47.417000',
   'updated_at': '2024-06-12T08:05:50.497000',
   'description': None,
   'serving_names': [
     'GROCERYCUSTOMERGUID'
   ],
   'catalog_name': 'Grocery Dataset Tutorial'
 },
 <featurebyte.api.entity.Entity at 0x29886edc0>
 {
   'name': 'productgroup',
   'created_at': '2024-06-12T08:05:48.244000',
   'updated_at': '2024-06-12T08:05:51.678000',
   'description': None,
   'serving_names': [
     'PRODUCTGROUP'
   ],
   'catalog_name': 'Grocery Dataset Tutorial'
 }]

In [16]:

# 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 [17]:

# Preview feature_group
feature_group.preview(preview_table)

# Preview feature_group feature_group.preview(preview_table)

Out[17]:

	POINT_IN_TIME	GROCERYINVOICEITEMGUID	CUSTOMER_x_PRODUCTGROUP_Time_Since_Latest_Timestamp	CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost_14d	CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost_28d	CUSTOMER_Latest_invoice_Amount	CUSTOMER_Count_of_invoice_14d	CUSTOMER_Count_of_invoice_28d	CUSTOMER_Avg_of_invoice_Amount_14d	CUSTOMER_Avg_of_invoice_Amount_28d	CUSTOMER_Std_of_invoice_Amount_14d	CUSTOMER_Std_of_invoice_Amount_28d
0	2023-05-28 19:27:14	15973b2f-2256-4caa-b65b-cbbfdff0905b	1106.468611	NaN	NaN	4.17	1.0	1	4.170	4.170000	0.000000	0.000000
1	2023-02-07 11:04:26	fd1caae1-77e6-4667-8c83-df13f05bf2f5	360.113056	NaN	8.28	36.35	1.0	3	36.350	43.086667	0.000000	30.821573
2	2023-03-31 18:50:00	213ef7d3-c27b-43e0-bc0a-57d6c7c254b0	269.623333	8.00	8.00	53.11	1.0	1	53.110	53.110000	0.000000	0.000000
3	2022-09-18 18:52:36	ac7edfb5-63ed-49fb-9b89-76b0288ed2f8	99.261667	16.63	24.13	95.05	6.0	12	34.825	31.634167	27.195796	26.904648
4	2022-12-26 15:01:07	264f79fd-c24a-47cc-8a68-fe3753a4d74b	73.213056	6.28	14.46	39.25	1.0	6	39.250	44.955000	0.000000	37.251150
5	2022-08-17 19:13:52	40a07ca4-a991-4d21-b5cf-74ee61220f96	343.933611	NaN	6.00	34.27	NaN	4	NaN	12.052500	NaN	13.394768
6	2023-04-11 17:23:57	6084f39f-9d2c-4111-b1cc-502e1559c0c0	197.481944	6.00	6.00	19.70	2.0	2	18.490	18.490000	1.210000	1.210000
7	2022-12-10 21:08:26	77d02174-f1e1-41c1-9fb9-01c6246b0009	NaN	NaN	NaN	31.73	1.0	1	31.730	31.730000	0.000000	0.000000
8	2023-03-17 11:15:09	1b627a25-7eb4-4f61-b243-c93db487bff0	140.312500	1.67	1.67	9.82	5.0	6	23.380	21.408333	21.979151	20.542799
9	2023-05-05 08:00:42	57ca0770-eb8b-4769-8e67-eb1b7cc0a934	1097.448611	NaN	NaN	8.47	2.0	4	14.795	11.632500	6.325000	6.224180

Save features into catalog¶

With feature groups we can do it in one call.

In [18]:

feature_group.save()

feature_group.save()

Done! |████████████████████████████████████████| 100% in 9.1s (0.11%/s)         
Done! |████████████████████████████████████████| 100% in 6.1s (0.17%/s)         
Loading Feature(s) |████████████████████████████████████████| 10/10 [100%] in 0.

Add description and see feature definition files¶

In [19]:

# Add description
customer_x_productgroup_time_since_latest_timestamp.update_description(
    "Time Since Latest interaction between the customer and the product group"
)
# See feature definition file
customer_x_productgroup_time_since_latest_timestamp.definition

# Add description customer_x_productgroup_time_since_latest_timestamp.update_description( "Time Since Latest interaction between the customer and the product group" ) # See feature definition file customer_x_productgroup_time_since_latest_timestamp.definition

Out[19]:

# Generated by SDK version: 1.1.0.dev7
from bson import ObjectId
from featurebyte import DimensionTable
from featurebyte import FeatureJobSetting
from featurebyte import ItemTable
from featurebyte.api.request_column import RequestColumn


# dimension_table name: "GROCERYPRODUCT"
dimension_table = DimensionTable.get_by_id(ObjectId("666956c78080c62d0dc616e2"))
dimension_view = dimension_table.get_view(
    view_mode="manual", drop_column_names=[], column_cleaning_operations=[]
)

# item_table name: "INVOICEITEMS", event_table name: "GROCERYINVOICE"
item_table = ItemTable.get_by_id(ObjectId("666956c58080c62d0dc616e1"))
item_view = item_table.get_view(
    event_suffix=None,
    view_mode="manual",
    drop_column_names=["record_available_at"],
    column_cleaning_operations=[],
    event_drop_column_names=["record_available_at"],
    event_column_cleaning_operations=[],
    event_join_column_names=[
        "Timestamp",
        "GroceryInvoiceGuid",
        "GroceryCustomerGuid",
        "tz_offset",
    ],
)
joined_view = item_view.join(
    dimension_view, on="GroceryProductGuid", how="left", rsuffix="", rprefix=""
)
grouped = joined_view.groupby(
    by_keys=["GroceryCustomerGuid", "ProductGroup"], category=None
).aggregate_over(
    value_column="Timestamp",
    method="latest",
    windows=[None],
    feature_names=["CUSTOMER_x_PRODUCTGROUP_Latest_Timestamp"],
    feature_job_setting=FeatureJobSetting(
        blind_spot="120s", period="3600s", offset="120s"
    ),
    skip_fill_na=True,
    offset=None,
)
feat = grouped["CUSTOMER_x_PRODUCTGROUP_Latest_Timestamp"]
request_col = RequestColumn.point_in_time()
feat_1 = (request_col - feat).dt.hour
feat_1.name = "CUSTOMER_x_PRODUCTGROUP_Time_Since_Latest_Timestamp"
output = feat_1
output.save(_id=ObjectId("6669575033eb5cd5aebc1fe9"))

In [20]:

# Add description
customer_productgroup_sum_of_totalcost_14d = \
customer_productgroup_sum_of_totalcost_14d_28d["CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost_14d"]
customer_productgroup_sum_of_totalcost_14d.update_description(
    "Total spent by the customer on the product group over a 14d period."
)
customer_productgroup_sum_of_totalcost_28d = \
customer_productgroup_sum_of_totalcost_14d_28d["CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost_28d"]
customer_productgroup_sum_of_totalcost_28d.update_description(
    "Total spent by the customer on the product group over a 28d period."
)
# See feature definition file
customer_productgroup_sum_of_totalcost_28d.definition

# Add description customer_productgroup_sum_of_totalcost_14d = \ customer_productgroup_sum_of_totalcost_14d_28d["CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost_14d"] customer_productgroup_sum_of_totalcost_14d.update_description( "Total spent by the customer on the product group over a 14d period." ) customer_productgroup_sum_of_totalcost_28d = \ customer_productgroup_sum_of_totalcost_14d_28d["CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost_28d"] customer_productgroup_sum_of_totalcost_28d.update_description( "Total spent by the customer on the product group over a 28d period." ) # See feature definition file customer_productgroup_sum_of_totalcost_28d.definition

Out[20]:

# Generated by SDK version: 1.1.0.dev7
from bson import ObjectId
from featurebyte import DimensionTable
from featurebyte import FeatureJobSetting
from featurebyte import ItemTable


# dimension_table name: "GROCERYPRODUCT"
dimension_table = DimensionTable.get_by_id(ObjectId("666956c78080c62d0dc616e2"))
dimension_view = dimension_table.get_view(
    view_mode="manual", drop_column_names=[], column_cleaning_operations=[]
)

# item_table name: "INVOICEITEMS", event_table name: "GROCERYINVOICE"
item_table = ItemTable.get_by_id(ObjectId("666956c58080c62d0dc616e1"))
item_view = item_table.get_view(
    event_suffix=None,
    view_mode="manual",
    drop_column_names=["record_available_at"],
    column_cleaning_operations=[],
    event_drop_column_names=["record_available_at"],
    event_column_cleaning_operations=[],
    event_join_column_names=[
        "Timestamp",
        "GroceryInvoiceGuid",
        "GroceryCustomerGuid",
        "tz_offset",
    ],
)
joined_view = item_view.join(
    dimension_view, on="GroceryProductGuid", how="left", rsuffix="", rprefix=""
)
grouped = joined_view.groupby(
    by_keys=["GroceryCustomerGuid", "ProductGroup"], category=None
).aggregate_over(
    value_column="TotalCost",
    method="sum",
    windows=["28d"],
    feature_names=["CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost_28d"],
    feature_job_setting=FeatureJobSetting(
        blind_spot="120s", period="3600s", offset="120s"
    ),
    skip_fill_na=True,
    offset=None,
)
feat = grouped["CUSTOMER_x_PRODUCTGROUP_Sum_of_item_TotalCost_28d"]
output = feat
output.save(_id=ObjectId("6669575033eb5cd5aebc1fe6"))

In [21]:

# Add description
customer_latest_invoice_amount.update_description("Latest invoice Amount for the customer")
# See feature definition file
customer_latest_invoice_amount.definition

# Add description customer_latest_invoice_amount.update_description("Latest invoice Amount for the customer") # See feature definition file customer_latest_invoice_amount.definition

Out[21]:

# Generated by SDK version: 1.1.0.dev7
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("666956c38080c62d0dc616e0"))
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", period="3600s", offset="120s"
    ),
    skip_fill_na=True,
    offset=None,
)
feat = grouped["CUSTOMER_Latest_invoice_Amount"]
output = feat
output.save(_id=ObjectId("6669575033eb5cd5aebc1fea"))

In [22]:

# Add description
customer_count_of_invoice_14d = customer_count_of_invoice_14d_28d["CUSTOMER_Count_of_invoice_14d"]
customer_count_of_invoice_14d.update_description(
    "Count of invoice for the customer over a 14d period."
)
customer_count_of_invoice_28d = customer_count_of_invoice_14d_28d["CUSTOMER_Count_of_invoice_28d"]
customer_count_of_invoice_28d.update_description(
    "Count of invoice for the customer over a 28d period."
)
# See feature definition file
customer_count_of_invoice_28d.definition

# Add description customer_count_of_invoice_14d = customer_count_of_invoice_14d_28d["CUSTOMER_Count_of_invoice_14d"] customer_count_of_invoice_14d.update_description( "Count of invoice for the customer over a 14d period." ) customer_count_of_invoice_28d = customer_count_of_invoice_14d_28d["CUSTOMER_Count_of_invoice_28d"] customer_count_of_invoice_28d.update_description( "Count of invoice for the customer over a 28d period." ) # See feature definition file customer_count_of_invoice_28d.definition

Out[22]:

# Generated by SDK version: 1.1.0.dev7
from bson import ObjectId
from featurebyte import EventTable
from featurebyte import FeatureJobSetting


# event_table name: "GROCERYINVOICE"
event_table = EventTable.get_by_id(ObjectId("666956c38080c62d0dc616e0"))
event_view = event_table.get_view(
    view_mode="manual",
    drop_column_names=["record_available_at"],
    column_cleaning_operations=[],
)
grouped = event_view.groupby(
    by_keys=["GroceryCustomerGuid"], category=None
).aggregate_over(
    value_column=None,
    method="count",
    windows=["28d"],
    feature_names=["CUSTOMER_Count_of_invoice_28d"],
    feature_job_setting=FeatureJobSetting(
        blind_spot="120s", period="3600s", offset="120s"
    ),
    skip_fill_na=True,
    offset=None,
)
feat = grouped["CUSTOMER_Count_of_invoice_28d"]
output = feat
output.save(_id=ObjectId("6669575033eb5cd5aebc1fec"))

In [23]:

# Add description
customer_avg_of_invoice_amount_14d = customer_avg_of_invoice_amount_14d_28d["CUSTOMER_Avg_of_invoice_Amount_14d"]
customer_avg_of_invoice_amount_14d.update_description(
    "Avg of invoice Amount for the customer over a 14d period."
)
customer_avg_of_invoice_amount_28d = customer_avg_of_invoice_amount_14d_28d["CUSTOMER_Avg_of_invoice_Amount_28d"]
customer_avg_of_invoice_amount_28d.update_description(
    "Avg of invoice Amount for the customer over a 28d period."
)
# See feature definition file
customer_avg_of_invoice_amount_28d.definition

# Add description customer_avg_of_invoice_amount_14d = customer_avg_of_invoice_amount_14d_28d["CUSTOMER_Avg_of_invoice_Amount_14d"] customer_avg_of_invoice_amount_14d.update_description( "Avg of invoice Amount for the customer over a 14d period." ) customer_avg_of_invoice_amount_28d = customer_avg_of_invoice_amount_14d_28d["CUSTOMER_Avg_of_invoice_Amount_28d"] customer_avg_of_invoice_amount_28d.update_description( "Avg of invoice Amount for the customer over a 28d period." ) # See feature definition file customer_avg_of_invoice_amount_28d.definition

Out[23]:

# Generated by SDK version: 1.1.0.dev7
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("666956c38080c62d0dc616e0"))
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="avg",
    windows=["28d"],
    feature_names=["CUSTOMER_Avg_of_invoice_Amount_28d"],
    feature_job_setting=FeatureJobSetting(
        blind_spot="120s", period="3600s", offset="120s"
    ),
    skip_fill_na=True,
    offset=None,
)
feat = grouped["CUSTOMER_Avg_of_invoice_Amount_28d"]
output = feat
output.save(_id=ObjectId("6669575033eb5cd5aebc1fee"))

In [24]:

# Add description
customer_std_of_invoice_amount_14d = customer_std_of_invoice_amount_14d_28d["CUSTOMER_Std_of_invoice_Amount_14d"]
customer_std_of_invoice_amount_14d.update_description(
    "Std of invoice Amount for the customer over a 14d period."
)
customer_std_of_invoice_amount_28d = customer_std_of_invoice_amount_14d_28d["CUSTOMER_Std_of_invoice_Amount_28d"]
customer_std_of_invoice_amount_28d.update_description(
    "Std of invoice Amount for the customer over a 28d period."
)

# See feature definition file
customer_std_of_invoice_amount_28d.definition

# Add description customer_std_of_invoice_amount_14d = customer_std_of_invoice_amount_14d_28d["CUSTOMER_Std_of_invoice_Amount_14d"] customer_std_of_invoice_amount_14d.update_description( "Std of invoice Amount for the customer over a 14d period." ) customer_std_of_invoice_amount_28d = customer_std_of_invoice_amount_14d_28d["CUSTOMER_Std_of_invoice_Amount_28d"] customer_std_of_invoice_amount_28d.update_description( "Std of invoice Amount for the customer over a 28d period." ) # See feature definition file customer_std_of_invoice_amount_28d.definition

Out[24]:

# Generated by SDK version: 1.1.0.dev7
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("666956c38080c62d0dc616e0"))
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="std",
    windows=["28d"],
    feature_names=["CUSTOMER_Std_of_invoice_Amount_28d"],
    feature_job_setting=FeatureJobSetting(
        blind_spot="120s", period="3600s", offset="120s"
    ),
    skip_fill_na=True,
    offset=None,
)
feat = grouped["CUSTOMER_Std_of_invoice_Amount_28d"]
output = feat
output.save(_id=ObjectId("6669575033eb5cd5aebc1ff0"))

Concepts in this tutorial¶

• Aggregate Features
• Window Aggregate Features

SDK reference for¶

• View.groupby()
• Groupby.aggregate_over()
• List of aggregate functions
• FeatureGroup
• FeatureGroup.save()