How EarlyNarratives works
Evidence-first signal tracking across public sources. This document explains the pipeline, the meaning of the dashboard metrics, and how to audit any signal back to sources.
No investment advice. Signals & sources only. EarlyNarratives provides informational signals derived from public sources. It does not provide financial, legal, or tax advice. Use at your own risk. We do not add new facts; we display summaries and metrics derived from collected posts and their metadata.
EarlyNarratives Whitepaper
Add your content to web/content/whitepaper.md.
What this is
EarlyNarratives is an evidence-first view of signals across public sources.
- Evidence-first: every signal should point to underlying sources (URLs + timestamps).
- Transparency over hype: show diversity, concentration, duplicates, and confidence signals.
- Multi-source confirmation matters: patterns across multiple publishers are stronger than single-source spikes.
- No recommendations: metrics describe attention and evidence structure, not what to do.
- Collect public posts from configured sources (e.g. news feeds and social platforms).
- Normalize and canonicalize URLs, extract origin domains, and enrich metadata.
- Deduplicate near-identical content to reduce amplification bias.
- Cluster related posts into signals and compute metrics per time window.
- Render weekly/daily briefings from the same payload shown in the UI.
Overall signal score for this signal in the selected window. Higher means more evidence/consistency, not a prediction.
Change in signal activity/score over the last 24 hours. Higher means accelerating attention, not performance.
Number of items included in the signal cluster for this window.
Unique origin domains/publishers (original sources) contributing to this signal. Higher = broader origin coverage.
Unique publishers/accounts observed (can include amplifiers and re-posters). Higher = broader publisher participation.
Share of near-duplicate items in the cluster. Higher can indicate repetition or amplification.
Concentration of origins: the top origin's share of items. Higher means fewer origins dominate the signal.
How many source types are represented (e.g., news vs social). Higher means more mixed evidence types.
Earliest post timestamp linked to a signal across all runs.
Latest post timestamp in the most recent snapshot for that signal.
Heuristic maturity/confidence score derived from breadth and consistency signals. Not a recommendation.
- Open the evidence drawer and click source links; verify publisher/domain diversity.
- Check concentration (top origin share) to see whether one outlet dominates.
- Check duplicate ratio to spot copy-paste waves and amplification.
- Use momentum to track acceleration; treat it as attention velocity, not outcome.
- Coverage depends on configured sources and can miss important context.
- Metrics are heuristic and can be noisy; always validate via sources.
- Summaries are constrained to observed content; they are not exhaustive reporting.
For a live example, open the signals dashboard.