Database architecture
Understand how Akashic separates build-time election research and modeling from production-serving data in Neon Postgres.
Akashic separates build-time research and modeling from production serving.
Build and research
The akashic-elections-v2 Neon project is the primary environment for ingestion, normalization, modeling, quality checks, and publication preparation. Governed analytics views route supported research queries.
Production serving
The little-queen-58462905 Neon project serves application-facing data and includes selected foreign-table links to build outputs. Its smaller schema is not a complete copy of the research system.
Reproducibility
A material database-derived claim should record:
- project and branch;
- database;
- observation or extraction time;
- query or generated artifact;
- geography and boundary vintage;
- source and evidence class;
- known limitations.
Documentation coverage
Database comments and structured schema documentation are incomplete and uneven. Treat a missing description as documentation debt, not proof that a field is safe to infer. Public field guarantees live in the data dictionary and API docs.
This page describes architecture. It does not expose credentials, private connection details, or internal datasets.