Signals Overview
Signals are quantitative classifications of market conditions, derived from fundamental data. They describe the state the market is in, not where price is going.
What is a Signal?
A Signal transforms raw fundamental data into a standardised, interpretable classification. Signals are designed to:
- Describe current conditions in terms a desk can act on
- Be comparable across time periods and markets
- Support systematic research and position sizing
Signal Structure
Each signal includes:
Column names vary by product, but every signal table follows the same shape.
Taking gbsi_us.sig_system_stress as the example:
| Field | Description |
|---|---|
signal_date | The date the signal describes |
gbsi_smoothed_score | The continuous composite score |
gbsi_smoothed_percentile | That score as a rolling percentile, which the bands are cut from |
stress_regime_label | The banded classification, for example TIGHT |
confidence_label | Confidence in the reading: HIGH / MODERATE / LOW |
str_effective_at | When the value became publicly knowable |
str_observed_at | When Snowtrail could first have held it |
str_ingested_at | When this row was written |
See each product's Signals page for its own columns, and the data dictionary for full definitions.
Signal Categories
Signals are organised by market and theme:
- Balance signals - Supply/demand balance indicators
- Storage signals - Inventory and storage-related metrics
- Flow signals - Trade flow and transportation indicators
- Weather signals - Weather-driven supply/demand impacts