HOW IT WORKS · METHODOLOGY
About Market Impact
AI-powered cryptocurrency market impact analysis — how dozens of crypto news sources become a single, readable signal.
At a glance
Articles analysed / day
Timeframes scored
Source outlets polled
Ingestion cadence
The platform continuously reads crypto news, scores each article for expected market impact, and groups related coverage into ranked narratives. Every live number you see traces back to a specific set of source articles with reasoning you can inspect.
HOW IT WORKS
Five stages, one signal pipeline
Ingestion
Continuously polling curated RSS feeds from tier-one outlets down to specialist publications.
Every new article is normalized (title, body, timestamp), deduplicated by content hash, and written to a structured store. Each source carries a credibility score used as a multiplier in downstream impact scoring.
AI evaluation
Each article evaluated on its own merits — headline, body, and metadata — scored across five timeframes and two market buckets.
The model scores each article across minute, hour, daily, weekly, monthly horizons and against BTC and ALT. The split reflects basic crypto market structure: Bitcoin dominates market capitalization, and altcoins collectively trade as a fraction of BTC. Plain-English "Why it matters" and "Expected impact" paragraphs are written for every article.
Statistical aggregation
Individual predictions don't mean much in isolation. A time-weighted model turns them into a coherent signal.
Each article's influence is strongest while the story is fresh and fades on a schedule that respects each timeframe's natural lifespan — a minute-scale reaction decays over minutes; a weekly narrative decays over weeks.
Narrative clustering
A second AI pass groups related articles into narratives — so you read each story once, not ten times.
Every narrative gets its own "Why it matters" and "Expected impact" summary derived from the full set of articles it contains, plus the underlying source list for transparency.
Contextual labeling
Is this high or low? Raw numbers are meaningless without a baseline.
Rolling percentiles over 7 and 30 days label every live value against its own history: very low → normal → very high. That context is what makes signals actionable rather than decorative.
METHODOLOGY & TEAM
Reproducible by design, built by one trader-developer
Methodology
Market Impact applies a statistical model to AI-scored news events. Every prediction is reproducible: given the same article, source credibility, and rolling baseline, the system produces the same signal. We deliberately avoid hand-tuned rules and per-market overrides.
Data sources
A curated set of English-language crypto news publishers via public RSS feeds. The list is reviewed periodically — underperforming sources removed, new ones added as the ecosystem evolves. No paywalls, no paid APIs, no content modification.
Team
Built and maintained by Gabriele Peloso — software engineer with a mathematics degree, 10+ years of mobile development, and 5+ years trading crypto. The project started as an experiment in applying statistical reasoning to the flood of crypto news.
What this site does not do
No trading advice. No price forecasts. The "impact" metric describes how much a news event's expected effect deviates from recent baselines — an analytical lens, not a buy/sell signal. Use it to understand what's happening, not to act on it blindly.
KEY FEATURES
What the platform gives you
FREQUENTLY ASKED
Quick answers
What is Market Impact?+
How does Market Impact evaluate articles?+
What are narratives?+
What news sources does Market Impact analyze?+
What timeframes are predictions available for?+
Is Market Impact free to use?+
How often is the data updated?+
Get in touch