Free beta. Start without signing in; sign in to save. Existing drafts stay intact.
Public beta · Researcher judgment required. Limits
REAL ARTIFACT · EXPLORATORY INVENTORYEPA 2025 PM2.5: 2,363 daily summary rows across 6 sitesArchived September 22, 2026 · Atlanta-area reference inventoryArtifact details & limits
Demonstrates: an inspectable exploratory inventory using identified real EPA data.
Does not establish: a causal finding, independent validation, or proof that ASTRA improves research outcomes. This artifact is separate from the fictional heat-loss walkthrough.
How should we measure reporting-day availability around Atlanta? Raw EPA row counts can mislead because sites report multiple measurement streams—up to 3.3 rows per reported day in this inventory. We propose distinct reporting days per site and per stream, with no site ranking.
Scope and verification status
The inventory covers sites within 40 km of central Atlanta, across five counties. A stream is identified by POC, method code and sample duration. These are reporting-day counts, not a completeness or population-coverage measure.
A developer-run checksum and count check passed on the retained CSV. The original EPA archive was not reverified.
Status: draft. Challenges are AI-generated; no human reviewer has responded and nothing has been independently reproduced. This record shows reasoning, not a finding about air quality.
ASTRA is a research workspace today; a wider scientific community is still being built. Our ambition is to support the scientific cooperation needed for a Type I civilization. The practical starting point is a research decision that others can inspect, challenge and improve.
Limits
What the platform can establish today
ASTRA has no completed, independently reviewed study demonstrating its effectiveness. Source-change flags appear in linked decision records when you load or refresh them after a change is reported. Selected Crossref DOIs and topics support scheduled checks; coverage depends on the configured watchlist and connected runner. This is not comprehensive literature monitoring. Recorded objections and revised plans do not establish scientific validity.
Account saving requires sign-in. Saving does not share a project; reviewer access requires a separate action. Review history provides tamper evidence, not an immutable ledger. Pilot pricing and support require agreement; no checkout is enabled.