Last Updated: caa. 20th Jan 22
Following JSON details the fields that will be sent to both Slacker Manager. The source for the intrabank transactions historical features is retrieved from the following source.
transactions.transactions_interbank_transaction_occurred_event_v3
BigQuery Table:
PROJECT_ENV.transactions.transactions_interbank_transaction_occurred_event_v3
Current method of retrieving the historical features, is to run an SQL query against the specified BigQuery table and aggregate the historical features as part of this query. A sample of the query execution will return the following table and values.
The keys of the table are:
cust_id
-senderInfo.customerId
beneficiary_acc_id
-interbankBeneficiaryDetails.accountNumber
Historical features are calculated as the difference in the number of hours between now and the expectedDate of transaction. Here one day is defined as 24 hours in the hour_diff
variable.
DATETIME_DIFF(CURRENT_DATE(), SAFE_CAST(expectedDate as date), HOUR) as hour_diff
360 days:
hour_diff <= 360*24
30 days:
hour_diff <= 30*24
7 days:
hour_diff <= 7*24
1 day:
hour_diff <= 1*24
Interbank Transaction Features
transaction_interbank_history_features
Response Result | Description | feature_id_prefix | feature_id | Type |
---|---|---|---|---|
customerId |
|
|
| UUID |
unique_tx_days | Number of distinct dates in which transactions occured |
|
| INT |
transactionType | The type of transaction as defined under transactions squad and anti-fraud rules |
|
| STRING |
category | The category under the corresponding transaction type |
|
| STRING |
direction | The direction of the transaction
|
|
| STRING |
status | The status of the transaction - E.g. PENDING, SUCCESS, etc. |
|
| STRING |
incoming_unique_days_360d | Number of distinct dates in which incoming transactions occurred in the last 360 days |
|
| INT |
incoming_avg_amt_360d | SUM(incoming transactions in the last 360 days) / incoming_unique_days_360d |
|
| FLOAT |
outgoing_unique_days_360d | Number of distinct dates in which outgoing transactions occurred in the last 360 days |
|
| INT |
outgoing_avg_amt_360d | SUM(outgoing transactions in the last 360 days) / outgoing_unique_days_360d |
|
| FLOAT |
tx_total_count | Total count of transactions made till date |
|
| INT |
tx_count_1d | Total count of transactions made in the last 24 hours |
|
| INT |
tx_sum_1d | Total amount of transactions made in the last 1 day |
|
| FLOAT |
tx_count_7d | Total count of transactions made in the last 7 days |
|
| INT |
tx_sum_7d | Total amount of transactions made in the last 7 days |
|
| FLOAT |
tx_count_30d | Total count of transactions made in the last 30 days |
|
| INT |
tx_sum_30d | Total amount of transactions made in the last 30 days |
|
| FLOAT |
tx_count_360d | Total count of transactions made in the last 360 days |
|
| INT |
tx_sum_360d | Total amount of transactions made in the last 360 days |
|
| FLOAT |
tx_sum_last5_inc | Sum of the last 5 transactions that occurred |
|
| FLOAT |