What a banking data ontology looks like

A banking data ontology names the core concepts a bank works with, such as party, relationship, account, product, balance, GL account, legal entity and branch, and defines how they connect. A party holds accounts, each account is an instance of a product, and each posts to GL accounts. With those links written down, a question can cross systems without guessing.

The example below is a starting point, not a standard. Adjust it to how your bank works.

Core concepts

  1. Party. A person or organization the bank deals with, such as a customer or guarantor.
  2. Relationship. A group of parties managed together, such as a commercial household.
  3. Account. A deposit, loan or other contract the bank holds with a party.
  4. Product. The type of account, such as interest checking or commercial real estate loan.
  5. Balance. An amount on an account at a point in time: ending, average or available.
  6. GL account. Where the account's activity posts in the general ledger.
  7. Legal entity. The bank or subsidiary that books the account.
  8. Branch or cost center. Where the account is managed and measured.
  9. Collateral. An asset securing a loan.

Key relationships

  1. A party belongs to a relationship.
  2. A party holds one or more accounts.
  3. Each account is an instance of one product.
  4. Each account has balances over time.
  5. Each account posts to GL accounts and is booked at one legal entity and one branch.
  6. A loan account can be secured by collateral.

One question, traced

Commercial deposit growth by relationship starts at the relationship, finds its parties, finds their deposit accounts, filters by commercial deposit products, and takes average balances for the quarter. Each step depends on a definition someone has to approve: which products count as commercial deposits, and which balance type the board sees.

Where banks disagree internally

  1. What counts as a relationship.
  2. Which products are commercial.
  3. Which balance type a report uses.

How Preql helps

Preql's agents find how each of your systems represents the same concepts, from parties to products to balances. Business owners approve the mappings, and every change is versioned with lineage back to source.