Data Quality & Parser Review

Ensuring Transaction Accuracy

The Parser Review system is the foundation of Credit Card Co-Pilot’s intelligence. While the app automatically ingests data from SMS and Email, financial notifications from banks are often cryptic or incomplete. The Parser Review interface empowers you to audit, refine, and validate your transaction history, ensuring that your reward recommendations and milestone tracking remain 100% accurate.

The Parser Review Dashboard

Located within the app's management tools, the Review screen provides a centralized view of every data point captured by the system. It categorizes incoming data into four distinct states to help you focus on what needs attention:

• Ambiguous: Transactions where the merchant or category wasn't identified with high confidence. These are the highest priority for review.
• Parsed: Successfully identified transactions that are already contributing to your reward analytics.
• Failed: Data points that couldn't be structured (e.g., a non-financial SMS or a highly unusual bank format).
• All: A complete chronological audit log of all ingestion attempts.

Refining Your Data

When the system is unsure about a transaction, you can step in to provide the missing context. This human-in-the-loop approach ensures that "messy" bank data never breaks your financial insights.

Key Editable Attributes:

• Merchant Label: Rename cryptic merchant IDs (e.g., AMZN_MKTP_IND) to friendly names (Amazon).
• Amount & Currency: Verify the final transaction value to ensure milestone progress is tracked correctly.
• Transaction Channel: Toggle between Online and In-Store (POS). Since many credit cards offer different reward rates based on the channel, this distinction is critical for maximizing points.
• Event Kind: Reclassify entries as Transactions, Statement Credits, or Promotional Bonuses.

Intelligent Feedback Loop

Every time you edit or verify a transaction, the system records the change in a versioned metadata history. This not only keeps your current record clean but also helps the underlying engine learn from your preferences.

// Example of a refined ingestion event
{
  "merchantLabel": "Starbucks",
  "channel": "pos",
  "parseStatus": "parsed",
  "metadata": {
    "reviewVersion": 2,
    "reviewLastAction": "edit",
    "notes": "Corrected merchant name from STBKS-MUM"
  }
}

Why Data Quality Matters

The precision of your data directly impacts the value you get from the app:

1. Optimal Recommendations: The recommendation engine uses your historical merchant data to predict where you spend most frequently. Clean merchant labels result in more accurate "Best Card" suggestions.
2. Milestone Precision: Many cards require specific spend thresholds (e.g., "Spend ₹1 Lakh for a ₹2,000 voucher"). Verified transaction amounts ensure you never miss a reward due to a parsing error.
3. Audit-Ready History: By resolving ambiguous events, you maintain a pristine digital passbook that accurately reflects your spending habits across all linked cards.

Quality Metrics Tracking

The system monitors its own performance through Ingestion Quality Metrics. By reviewing a 30-day window of ingestion events, the app provides a transparency report on how much of your data was captured perfectly versus how much required manual intervention. This allows the Co-Pilot to constantly improve its parsing logic for your specific banking providers.