Overview — Data access and delivery

A model that cannot ingest clean data or cannot be imported from another script is a demonstration, not a deliverable. This topic closes the programme: deepen API ingestion into a time-indexed frame without hard-coded secrets, treat SQL as the system of record for prices, package the pricer as an importable module, and run an applied project that loads a panel, estimates risk inputs, prices one instrument and tests an identity.

By the end of this note you should be able to:

• parse an API-like payload into a validated DatetimeIndex frame and handle missing fields;
• design a small SQLite price schema and query a panel from Python;
• lay out a minimal package and import the pricer from a separate script;
• execute the capstone workflow and change one input to recompute.

1. Contents
2. The big picture