Government sources
Official datasets and agency price monitoring provide authoritative reference points for commodities and fuel.
Data Sources & Methodology
Where our intelligence comes from
We believe trustworthy intelligence requires transparent methodology. This page documents our sources, collection and verification methods, and how confidence is scored.
Powered by Pulse Intelligence Engine
Source categories
Supplier sources
Verified suppliers submit prices, inventory and availability directly through the Supplier Portal.
Community sources
NeighborGrid contributors report on-the-ground prices and conditions, displayed in aggregate.
Future API sources (Coming Soon)
Direct API integrations and automated price feeds are planned to expand coverage and freshness.
Data collection methods
- Structured submissions through the Supplier Portal, stored historically with timestamps.
- Community signals collected through the NeighborGrid Network.
- Periodic imports from government and official datasets.
- Planned automated feeds and API integrations for real-time coverage.
Verification methods
Source verification
Suppliers and sources are assigned trust levels based on identity, history and corroboration.
Cross-corroboration
Community signals gain confidence as independent reports agree.
Moderation
Flagged or anomalous data is reviewed before it influences intelligence.
Confidence scores
Every data point carries a confidence score from 0–100, reflecting source reliability, recency and corroboration. Scores map to bands: high (90+), good (75–89), fair (60–74) and low (below 60).
Update frequencies
Commodity and fuel signals refresh continuously as new submissions and imports arrive. Each record displays its own last-updated timestamp.
Pulse insight methodology
The Pulse Intelligence Engine applies transparent, rule-based logic to detect patterns — price volatility, below-average pricing, rising fuel costs, growing demand and declining supply. Each insight is labelled with a signal strength and confidence level. Pulse does not currently perform statistical forecasting; outputs are clearly marked as data-driven insights or forecast previews.
Data quality principles
- Never overwrite records — all submissions are stored historically with timestamps.
- Always attribute sources and surface confidence.
- Prefer corroborated, verified data over single unverified signals.
- Label every estimate; never present an estimate as certainty.
Source attribution examples
| Example source | Trust level | Confidence |
|---|---|---|
| Government dataset | Verified Official | 95 |
| Verified supplier submission | Verified Supplier | 88 |
| Community report (3+ corroborations) | Community Verified | 74 |
| Single community report | Unverified | 52 |