BankOrg Docs

In progress

BankOrg Data Platform / Source Data / Generate Data

Initial Data

Generate the initial load of the core banking feeds: the snapshot of 2026-10-01, with transactions from 2026-09-01 to 2026-10-01.

The initial load is the snapshot of 2026-10-01. It is written by one Python script in the scripts/ folder of the repository, using only the standard library. All data is synthetic.

Date range

The dates inside the files run as follows:

  • branches: OPEN_DATE from 2004-06-01 to 2023-12-11.
  • customers: OPEN_DATE from 2008-01-16 to 2026-09-12; DATE_OF_BIRTH from 1955-02-20 to 2004-08-09.
  • accounts: OPEN_DATE from 2008-07-01 to 2026-09-30; CLOSE_DATE from 2026-07-02 to 2026-09-30 (35 closed accounts).
  • transactions: TRANSACTION_TS from 2026-09-01 00:32:20 to 2026-10-01 23:59:15, which is 31 days of history.

Datasets

  • branches: 50 branches in 21 provinces of Vietnam, opened between 2004-06-01 and 2023-12-11. Region: 18 SOUTH, 17 NORTH, 15 CENTRAL. Type: 1 HEAD_OFFICE, 34 FULL_SERVICE, 15 SATELLITE. Status: 48 ACTIVE, 2 CLOSED. Customers and accounts only use the active branches.
  • customers: 300 customers, each with a home branch among the active branches.
  • accounts: 494 accounts of type CHECKING, SAVINGS, LOAN or CREDIT_CARD, mostly in the customer's home branch.
  • transactions: 5074 rows from 2026-09-01 to 2026-10-01, all CHANGE_TYPE = I. They are 4271 transactions; 803 of them have a second row, the receiving side of a transfer between two accounts in the same currency.

Posting legs

The transactions file has two columns after TRANSACTION_ID: LEG_NO, always 1 in the initial load, and DIRECTION, CR for DEPOSIT, TRANSFER_IN and INTEREST and DR for the other types. For a transfer, the script also writes the other side: same TRANSACTION_ID, LEG_NO and amount, the opposite DIRECTION and type, and the counterparty account as the account. The first run of the script wrote 4410 rows for the snapshot of 2026-10-18, without these columns and rows; the file was generated again when they were added, and again when the snapshot date was moved to 2026-10-01 so that the daily dates start on 2026-10-02.

Generate

Terminal
python3 scripts/generate_initial_data.py
Output
data/initial/core_banking/branches/branches_original_50.csv: 50 rows (read only)
data/initial/core_banking/customers/customers_original_300.csv: 300 rows
data/initial/core_banking/accounts/accounts_original_494.csv: 494 rows
data/initial/core_banking/transactions/transactions_original_5074.csv: 5074 rows

Each dataset folder holds the CSV and a marker file _SUCCESS.json with the file name and the number of data rows.

Output of ls -ltr on the initial dataset folders, first run.