View a markdown version of this page

Forecasting data in the Connect Customer analytics data lake - Amazon Connect Customer

Forecasting data in the Connect Customer analytics data lake

This topic details the content in the Connect Customer data lake forecasting tables. Each table lists the column, type, and description of the content in the table.

There are two ways to access the analytics data lake and configure data to be shared:

If you are unable to access the scheduling tables by using Option 1, try using Option 2.

Important things to know

  • You can use the tables described in this topic to access published forecasts data in the data lake.

  • The Forecast groups table stores versioned records. A new version is created when forecast group details are changed, for example, adding or removing queues from the forecast group. You can get the latest record using the highest value of forecast_group_version.

  • You can join the Forecast groups table to the Long-term and Short-term forecasts tables by using the following columns: forecast_group_arn and forecast_group_version.

Forecast groups table

Table name: forecast_groups

Description: Defines forecast groups that organize queues and channels for demand forecasting, using versioning for change tracking.

Primary key: instance_id, forecast_group_arn, forecast_group_version

Join keys:

  • instance_id — Joins to all tables

  • forecast_group_arn, forecast_group_version — Joins to long_term_forecasts, short_term_forecasts

  • forecast_group_arn — Joins to staffing_group_forecast_groups

Column Type Description
instance_id String No The identifier of the Connect Customer instance.
forecast_group_arn String No The ARN of the forecast group.
forecast_group_version Number No

The version of the forecast group. A new version is created every time a change is made to a forecast group, for example, addition of new queues.

forecast_group_name String Yes The name of the forecast group.
instance_arn String Yes The ARN of the Connect Customer instance.
is_deleted Boolean Yes Whether the forecast group is deleted.
last_updated_timestamp String Yes The epoch Timestamp in milliseconds when the last time the forecast group was created/updated/deleted.
data_lake_last_processed_timestamp Timestamp Yes The Timestamp for the last time the data lake processed the record. This can include transformation and backfill processes. This field cannot be used to determine reliably data freshness.

Long-term forecasts table

Table name: long_term_forecasts

Description: Contains long-term (daily interval) forecast data including contact volume and average handle time predictions, with support for customer-applied overrides.

Primary key: instance_id, long_term_forcast_id

Join keys:

  • instance_id — Joins to all tables

  • forecast_group_arn, forecast_group_version — Joins to forecast_groups

  • queue_id — Joins to Agent Queue Statistic Record

Column Type Description
instance_id String No The ID of the Connect Customer instance

long_term_forcast_id

String Yes Unique Identifier of the forecast data row. Key is hash of multiple values: instanceId, forecastGroupId, forecastGroupVersion, forecastType, queueId, channel, forecastStarttime, creationTime.

forecast_group_arn

String Yes The ARN of the forecast group.
forecast_group_version Number Yes The version of the forecast group.
interval String Yes Time interval of the forecast data. For example, Daily for long term forecast data.
queue_id String Yes The ID of the queue for the forecast.
channel String Yes The channel of the forecast. For example, VOICE.
forecast_interval_start_time_ms Timestamp Yes Epoch in milliseconds of the start time of the time interval for this data row.
creation_timestamp_ms Timestamp Yes Epoch in milliseconds of when this forecast is first computed or published.
computed_timestamp_ms Timestamp Yes Epoch in milliseconds of when this forecast is first computed.
published_timestamp_ms Timestamp Yes Epoch in milliseconds of when this forecast is first published.
timezone String Yes The timezone of the forecast, for example, UTC.
is_published Boolean Yes Whether this forecast is published or not.
average_handle_time Number Yes The average handle time metric value of the forecast data row.
contact_volume Number Yes The contact volume metric value of the forecast data row.
average_handle_time_override Number Yes The customer applied override value of the average handle time metric.
contact_volume_override Number Yes The customer applied override value of the contact volume metric value.
instance_arn String Yes The ARN of the Connect Customer instance of the forecast.
data_lake_last_processed_timestamp Timestamp Yes The Timestamp for the last time the data lake processed the record. This can include transformation and backfill processes. This field cannot be used to determine reliably data freshness.

Short-term forecasts table

Table name: short_term_forecasts

Description: Contains short-term (15-minute interval) forecast data including contact volume and average handle time predictions, with support for customer-applied overrides.

Primary key: instance_id, short_term_forecast_id

Join keys:

  • instance_id — Joins to all tables

  • forecast_group_arn, forecast_group_version — Joins to forecast_groups

  • queue_id — Joins to Agent Queue Statistic Record

Column Type Description
instance_id String No The ID of the Connect Customer instance.

short_term_forecast_id

String Yes Unique Identifier of the forecast data row. Key is hash of multiple values: instanceId, forecastGroupId, forecastGroupVersion, forecastType, queueId, channel, forecastStarttime, creationTime.
forecast_group_arn String Yes The ARN of the forecast group for the forecast data row.
forecast_group_version Number Yes The version of the forecast group.
interval String Yes Time interval of the forecast data row. For example, FIFTEEN_MINUTES for short term 15 minutes forecast data row.
queue_id String Yes The ID of the queue for the forecast.
channel String Yes The channel of this forecast, for example, VOICE.
forecast_interval_start_time_ms Timestamp Yes Epoch in milliseconds of the start time of the time interval for this data row.
creation_timestamp_ms Timestamp Yes Epoch in milliseconds of when this forecast is first computed or published.
computed_timestamp_ms Timestamp Yes Epoch in milliseconds of when this forecast is first computed.
published_timestamp_ms Timestamp Yes Epoch in milliseconds of when this forecast is first published.
is_published Boolean Yes Whether this forecast is published or not.
average_handle_time Number Yes The average handle time metric value of the forecast data row.
contact_volume Number Yes The contact volume metric value of the forecast data row.
average_handle_time_override Number Yes The customer applied override value of the average handle time metric.
contact_volume_override Number Yes The customer applied override value of the contact volume metric value.
instance_arn String Yes The ARN of the Connect Customer instance of the forecast.
data_lake_last_processed_timestamp Timestamp Yes The Timestamp for the last time the data lake processed the record. This can include transformation and backfill processes. This field cannot be used to determine reliably data freshness.

Intraday forecasts table

Table name: intraday_forecasts

Description: Contains intraday forecast data including forecasted contact volume, average handle time, queue answer time, and effective agent staffing for real-time workforce adjustments.

Primary key: instance_id, intraday_forecast_id

Partition key: forecast_interval_start_timestamp (daily)

Join keys:

  • instance_id — Joins to all tables

  • queue_arn — Joins to Agent Queue Statistic Record (via queue ARN and ID mapping)

Column Type Nullable Description
intraday_forecast_id string No Unique identifier of this intraday forecast data.
aws_account_id string Yes The identifier of the AWS account that owns the Intraday Forecast.
instance_id string No The identifier of the Connect Customer instance. You can find the instance ID in the Amazon Resource Name (ARN) of the instance.
instance_arn string Yes Instance ARN of the Connect Customer instance.
channel string Yes The method used to contact your contact center.
queue_arn string Yes The Amazon Resource Name of the queue.
forecast_interval_start_time Timestamp Yes Start Timestamp of the forecast interval.
creation_timestamp Timestamp Yes When the forecast was computed in forecasting system.
average_handle_time Double Yes Forecasted metric data: average handle time.
average_queue_answer_time Double Yes Forecasted metric data: average queue answer time.
contact_volume Double Yes Forecasted metric data: contact volume.
effective_agent_staffing Double Yes Forecasted metric data: effective agent staffing.
data_lake_last_processed_timestamp Timestamp Yes Timestamp, which shows the last time the data lake processed the record. This can include transformation and backfill. This field cannot reliably be used to determine data freshness.

Demand group table

Table name: demand_group

Description: Defines demand groups that represent specific queue-channel combinations for capacity planning, associating demand targets with forecast groups.

Primary key: instance_id, demand_group_arn, demand_group_version

Join keys:

  • instance_id — Joins to all tables

  • foecast_group_arn — Joins to forecast_groups (as forecast_group_arn)

  • demand_group_arn, demand_group_version — Joins to demand_group_definitions

  • demand_group_arn — Joins to staffing_group_demand_group, staff_demand_group

Column Type Description
instance_id string No The identifier of the Connect Customer instance.
demand_group_arn string No The ARN of the demand group.
demand_group_version Long No

The version of the demand group. A new version is created every time a change is made to a demand group, for example, addition of new queues.

instance_arn string Yes The ARN of the Connect Customer instance.
demand_group_name string Yes Name of the demand group.
foecast_group_arn string Yes The ARN of the forecast group.
is_deleted Boolean Yes Whether the demand group is deleted.
last_updated_timestamp Timestamp Yes Timestamp when the demand group was last created/updated/deleted.
data_lake_last_processed_timestamp Timestamp Yes Timestamp, which shows the last time the record was touched by the data lake. This can include transformation and backfill. This field cannot reliably be used to determine data freshness.

Demand group definitions table

Table name: demand_group_definitions

Description: Defines the queue and channel combinations that make up each demand group, mapping specific workload types to demand allocation targets.

Primary key: instance_id, demand_group_definition_id

Join keys:

  • instance_id — Joins to all tables

  • demand_group_arn, demand_group_version — Joins to demand_group

  • queue_id — Joins to Agent Queue Statistic Record

Column Type Description
instance_id string No The identifier of the Connect Customer instance.
demand_group_definition_id string No Unique Identifier for the demandGroup definition row.
demand_group_arn string Yes The ARN of the demand group.
demand_group_version Long Yes

The version of the demand group. A new version is created every time a change is made to a demand group, for example, addition of new queues.

instance_arn string Yes The ARN of the Connect Customer instance.
queue_id string Yes ID of the queue that is part of the demand group.
is_deleted Boolean Yes Whether the demand group is deleted.
last_updated_timestamp Timestamp Yes Timestamp when the demand group was last created/updated/deleted.
data_lake_last_processed_timestamp Timestamp Yes Timestamp, which shows the last time the record was touched by the data lake. This can include transformation and backfill. This field cannot reliably be used to determine data freshness.