Logging Transactions

The Transaction Ledger is the primary input interface. Everything else in the platform — your portfolio performance, your holdings breakdown, the analytics — is computed from your transaction history.

Fields

Ticker

The stock or asset symbol. Must be a valid ticker recognisable by Yahoo Finance (e.g. AAPL, MSFT, BTC-USD, GC=F for gold futures).

If you enter a ticker the system hasn’t seen before, it will:

  1. Validate it against Yahoo Finance
  2. Backfill the full available price history into the database
  3. Then write your transaction

This means the first transaction for a new ticker takes a few extra seconds. Subsequent transactions for that ticker are instant.

Transaction Type

BUY or SELL.

  • BUY increases your position.
  • SELL decreases it. You can’t SELL more than you currently hold — the system will reject that.

Quantity (shares / units)

The number of shares or units transacted. Fractional shares are supported (e.g. 0.5 for half a Bitcoin).

Price per unit (USD)

The price at which you transacted, in USD. This is the price you paid or received, not the market price at time of submission. Enter your actual cost basis.

Transaction Date

The date the transaction occurred. Defaults to today’s date. You can backdate transactions freely — this is useful when onboarding and entering your historical trade history.

The Transaction Ledger table

The main panel shows all your transactions sorted by date, most recent first. Each row shows ticker, type, quantity, price, date, and total value.

Correcting Mistakes

If you entered a transaction incorrectly, use the Delete a transaction section at the bottom of the ledger panel. Enter the transaction ID (shown in the ledger table) and confirm deletion.

Deletion is intended for corrections only — there’s no audit trail suppression, so your models will recompute correctly after a delete.

After deleting, the pipeline will recompute your holdings on the next sensor tick or scheduled run.

What happens after you submit

  1. The transaction is written to bronze.transactions.
  2. The transaction_sensor in Dagster detects the new record.
  3. If the ticker is new, price history is backfilled.
  4. The dbt models downstream of stg_transactions are rerun: int_daily_holdings, fct_daily_returns, and all gold marts.
  5. Your analytics dashboard reflects the updated data.

Depending on when you submit, this can take anywhere from a few seconds (sensor tick) to a few minutes if there’s a backfill happening.