P&L Automation from Chats & Banks

I need an end-to-end workflow that turns two very different data sources—my chat feed and my bank statements—into a single, reliable profit-and-loss report.

Data sources
• Chat data feed: format is completely flexible; if you have a preferred structure (JSON, CSV, database, etc.) let me know and we’ll use whatever keeps the parsing clean.
• Monthly bank statements.

Key tasks
• Parse the chat feed and accurately pull only the revenue and expense lines.
• Combine those findings with the matching transactions in the bank statement.
• Reconcile any discrepancies and flag items that appear in one source but not the other.
• Produce a simple P&L summary (ideally both on-screen and downloadable as CSV or Excel).

Deliverables
1. Well-commented code or repeatable workflow (Python with pandas / SQL / similar) that performs the extraction, matching, and report generation.
2. A short README explaining setup, expected input folders, and how to run the process.
3. Sample output using a small dummy data set so I can see the layout before feeding live data.

Acceptance criteria
• All revenue and expense figures from both sources appear once—and only once—in the final report.
• Reconciliation log lists every unmatched line item.
• Script runs locally on macOS or in a lightweight cloud notebook with no paid dependencies.

If you already have experience parsing conversational data or reconciling multiple ledgers, please include a quick note about it in your proposal.