Where AlphaPulse data
comes from
Every AlphaPulse signal is derived entirely from public text — no price feeds, no paywalled sources. Here is how raw internet narrative becomes a structured market signal, and how you can access it.
Sources
Reuters, Bloomberg, market commentary sites, social platforms, analyst blogs, SEC filings, and domain-specific forums — all publicly available.
Collection & processing
Text is ingested every ~15 minutes. Domain-specific NLP models score each article for sentiment and topic relevance. Outlier filters and quality gates run before any score is published.
Output metrics
Two scores per topic per day: Pulse Strength (directional sentiment, −100 to +100) and Attention Pulse (coverage volume, 0–100), plus 1d / 7d / 30d change deltas.
AlphaPulse covers 100+ topics across equities, crypto, macro, monetary policy, inflation, forex, commodities, geopolitics, consumer, real estate, bonds, and thematic domains. Signals are purely narrative-derived — they complement, not replace, traditional price data.
Frequently asked questions
- What exactly are AlphaPulse signals?
- Each signal is a pair of scores: Pulse Strength (a directional sentiment score from −100 to +100) and Attention Pulse (a coverage/volume score from 0 to 100). Together they tell you whether a topic is being discussed positively or negatively, and how loudly.
- Where does the underlying data come from?
- We continuously ingest publicly available text — financial news wires, market commentary, social platforms, analyst blogs, earnings transcript coverage, regulatory filing discussions, and specialised forums. We do not use proprietary price feeds or paywalled data sources.
- Is this price data or sentiment data?
- Purely sentiment and attention — no OHLCV price data is included. AlphaPulse signals are leading or coincident indicators derived from narrative flow, not from market prices themselves. They are designed to complement, not replace, traditional market data.
- Can I use AlphaPulse data in my own models or systems?
- Yes — that is exactly what the Pro API tier is for. The REST API returns JSON signal data that you can pull into any quant pipeline, spreadsheet, or internal dashboard. Contact us if you need bulk historical exports or custom delivery formats.
Further reading
Investor Sentiment and the Cross-Section of Stock Returns
Baker & Wurgler · Journal of Finance, 2006
The foundational study establishing that investor sentiment — measured via observable proxies — systematically predicts future stock returns across market segments.
Twitter Mood Predicts the Stock Market
Bollen, Mao & Zeng · arXiv, 2010
Landmark research demonstrating that collective mood states extracted from Twitter feeds correlate with and predict Dow Jones movements, achieving 87.6% directional accuracy.
Textual Analysis in Finance
Loughran & McDonald · Annual Review of Financial Economics, 2020
A comprehensive review of NLP methods applied to financial documents — earnings calls, news, filings — and their value as signals for asset pricing and risk assessment.
Request a data sample
Get 1 year of daily signal history for any tracked topic delivered to your inbox as CSV. One historical dataset request is free — no account required.
Access tiers
Full pricing →Free delay: 3 h (crypto, equities) to 24 h (macro, real estate, geopolitics).
Upgrade →