Data sources
Primary public data, catalogued, with refresh cadence and limitations.
Capitora ingests only public, attributable data. This page lists every source we currently use, what we use it for, how often we refresh it, and the known limitations of each. If a source is not listed here, we do not use it.
01 · Principles
Sourcing principles.
Three rules govern what we ingest and how we present it.
- Primary first. We prefer the original document (filing, award, permit) to any secondary reporting. Where secondary coverage is included, it is labeled as such.
- Attributable always. Every signal cites the source it was derived from. No source means no signal.
- Public and lawful. We do not ingest non-public information, scraped paywalled content, or data acquired in violation of the publishing party's terms.
02 · Federal & procurement
Federal obligations, contracts, and agency activity.
| Source | Use | Refresh | Notes |
|---|---|---|---|
| USAspending.gov | Federal obligation flow by agency, recipient, NAICS, place of performance | Daily | Authoritative for obligated dollars; outlays lag obligations |
| SAM.gov | Contract awards, opportunities, vendor registrations | Daily | Award detail occasionally redacted; we annotate when so |
| GSA Schedule 70 | Civilian IT and cyber procurement velocity | Weekly | Vendor share data normalized across reporting cycles |
| DoD contract announcements | Same-day visibility into > $7.5M defense awards | Daily (business) | Subset of SAM but earlier in publication cycle |
| Agency budget justifications | Forward-look on planned obligations by program | Annual + amendments | Used for context, not signal generation |
03 · Markets & filings
Securities filings and institutional positioning.
| Source | Use | Refresh | Notes |
|---|---|---|---|
| SEC EDGAR (8-K / 10-Q / 10-K) | Corporate capex, supplier disclosures, contract awards | Continuous | Material 8-Ks parsed on publication; quarterly filings on receipt |
| SEC Form 4 | Insider transactions in implicated names | Daily | Aggregated; we do not surface individual insiders |
| SEC 13F | Quarterly institutional positioning deltas | Quarterly (45-day lag) | Used directionally; staleness is reported on every signal |
| NYSE / NASDAQ reference data | Symbol, sector, market-cap banding | Daily | Qualitative banding only; we do not display intraday quotes |
04 · Energy & infrastructure
Grid, generation, and infrastructure filings.
| Source | Use | Refresh | Notes |
|---|---|---|---|
| FERC eLibrary & OASIS | Generator interconnection, transmission filings | Continuous | OASIS posting cadence varies by RTO |
| EIA-860 / EIA-861 / EIA-930 | Generation capacity, utility sales, hourly grid demand | Monthly / hourly | EIA-930 used for real-time load context |
| ERCOT / PJM / MISO / CAISO postings | Reliability council load forecasts, capacity auctions | Per ISO cadence | Used for state-momentum and grid-modernization signals |
| NRC docket filings | Reactor restarts, license amendments, SMR applications | Continuous | Used for nuclear-instrumentation signals |
| DOE program announcements | Loan Programs Office, Grid Deployment Office, CHIPS R&D | As published | Conditional commitments versus closings distinguished |
05 · State & permitting
State economic-development and permitting feeds.
| Source | Use | Refresh | Notes |
|---|---|---|---|
| State public utility commission dockets | IRPs, CPCN filings, rate cases | Continuous | Coverage prioritized for states with active capital flows |
| County building & development records | Data-center and fab permit square footage | Per county cadence | Loudoun, Maricopa, Williamson, Franklin, Travis prioritized |
| State air permitting databases | Emissions controls tied to industrial capex | Per state cadence | Used for second-order supplier signals |
| State economic-development announcements | Site selection, incentive packages | As published | Cross-checked against permitting and utility filings before use |
| Commerce CHIPS Program Office | Preliminary and final award memos | As published | Distinguished from disbursed funds |
06 · Research & reference
Reference, contextual, and academic sources.
| Source | Use | Notes |
|---|---|---|
| Bureau of Labor Statistics, Census, BEA | Macroeconomic context for state and sector framing | Reference only |
| Federal Reserve H.15, FRED | Rates and macro reference series | Reference only |
| Peer-reviewed academic literature | Methodology grounding (e.g. interconnection queue dynamics, supplier-graph traversal) | Cited inline in methodology |
| Industry reference reports | Reliability council load outlooks, transformer lead-time surveys | Cited; not republished |
07 · Limitations
Known limitations and what we do about them.
No data source is perfect. We disclose the most significant limitations of each layer and how Capitora's pipeline accounts for them.
- Obligation–outlay gap. USAspending obligations precede actual disbursement and may be deobligated. We tag obligation-driven signals as forward-looking and require corroborating filings before elevating severity.
- 13F staleness. 13F data is reported with a 45-day lag and excludes short positions. We use it directionally and never as a stand-alone signal.
- Permit-to-energization slippage. Data-center permits do not always energize on schedule. We cross-check with utility interconnection queue data before treating permits as binding.
- Redacted procurement detail. Some defense and intelligence awards are partially redacted. Signals drawn from redacted awards are flagged accordingly.
- Coverage breadth bias. Our county- and ISO-level coverage is deepest in jurisdictions with active capital flows. We disclose coverage gaps; we do not infer absence from absence.
Source corrections
If you identify an error in how we have characterized a source, write to intelligence@capitora.ai with the citation. Corrections are logged in the methodology changelog.