↓ Skip to main content
  1. Posts/

The Quant's Sourcebook: Open Financial News, Data & NLP Feeds That Still Work (2026 Edition)

Author
Yauheni Yakauleu
Short, practical articles on quant methods, alternative data, and applied AI. Code Insights, Touch IT LLC.

Every quant accumulates a pile of bookmarks for research: squawk feeds, economic calendars, 13F screeners, data APIs, NLP papers. This is mine — cleaned up, tested this week, and reorganized into what a buy-side research workflow actually consumes: newswires first, releases second, data third, text-mining last. Every link was checked at the time of writing; links that are paywalled, login-gated, or bot-gated are marked as such, and the ones that died are in a graveyard section at the end so you don’t waste an afternoon the way I once did.

1. Live newswires and squawk (free or near-free)
#

2. Economic releases, calendars and official data
#

3. Alternative data releases worth automating
#

The high-signal free releases that institutional desks quietly watch:

  • TSA passenger volumes — daily air-travel demand, a real-time read on discretionary spend and airline capacity.
  • ICSC — mall/retail traffic research from the shopping-center industry.
  • OCC market-data reports (bot-gated) — options volume and open interest downloads; the base for positioning in the options market.
  • FBI NICS background checks (fbi.gov/cjis/nics, bot-gated) — the classic monthly alternative-data release; firearm-purchase checks proxy for certain consumer categories.
  • SEC EDGAR full-text search — machine-readable filings; note SEC asks scrapers to declare a descriptive User-Agent with contact info.
  • BlackRock Investment Institute weekly commentary and Morgan Stanley Ideas — free institutional thought flow, useful for reading what allocations desks are telling clients.

4. Price, fundamentals and quant data
#

  • Nasdaq Data Link — Quandl’s successor; freemium tabular and alternative datasets.
  • Polygon.io, Alpaca Data, Databento, Tiingo — the modern US equity-data APIs with usable free or low tiers. Databento is the closest to an institutional-grade tick store with a self-serve API.
  • QuantQuote and CSI Data — clean historical daily/intraday archives (CSI is a survivor of every cleanup; QuantQuote convenient for EOD backtests).
  • IQFeed — a real-time feed without a wall of licensing paperwork (pay per month, API-ready).
  • Kaggle datasets and AWS Registry of Open Data — for one-off research corpora.
  • Numerai — if you want your models tested against a live hedge-fund tournament data feed.
  • QuantConnect — the community successor to the platform listed in the graveyard below; free backtest compute, institutional-style data layers.
  • Yahoo Finance endpoints: the chart API still works but rate-limits aggressively (429s from anything but a browser session), and the legacy feeds.finance.yahoo.com RSS 2.0 endpoint went away (404 as of today). Cache locally and back off — don’t put Yahoo in a request loop.

5. Text-mining and NLP toolchain
#

  • NLTK — the book is still the fastest onboarding to tokenization, chunking, and information extraction.
  • spaCy — production NER pipelines; the base for any entity-aware news processor.
  • Hugging Face datasets — pre-built financial corpora (earnings calls, filings, headlines) that used to take weeks to assemble.
  • GDELT — global news-event database updated every 15 minutes; free, machine-readable, and the only realistic substitute for commercial event feeds.
  • Common Crawl / CC-NEWS — raw news at scale for training or backtesting on the historical record.
  • arXiv q-fin — the preprint firehose for quant research.

6. The pattern I actually use for news scanning
#

The single most reliable free structured feed is Google News RSS. Two implementation notes from my own pipeline, learned the hard way:

  1. Batch tickers with OR, not commas. In my feed cache this is a URL like https://news.google.com/rss/search?q=(AMZN OR NVDA OR COST) when:7d&hl=en-US&gl=US&ceid=US:en — roughly 100 usable items per batch, URL-encoded, one request per group. A comma-separated query looks the same but Google returns near-matches and junk, which is enough to send a naive pipeline into an infinite retry loop.
  2. Cache locally, then process. The digest step (headline + date + source) is where all the value is: my daily brief is generated from a local cache of nine batch feeds plus benchmarks, because relying on live endpoints — as of this week, Yahoo’s RSS is retired and their chart API throttles — turns infrastructure into a single point of failure.

7. Screening and flow tools
#

8. Professional portals (login-gated, but know they exist)
#

These shaped what “sell-side research” means and are worth requesting access to as your career allows: Goldman Sachs Research, UBS Neo, Citi Velocity, J.P. Morgan Markets, StreetAccount, Kensho — plus the Nomura, Morgan Stanley, Barclays and BNP research portals. Most are credential-walled; the free analogues above cover ~80% of the actual information content.

9. Modeling libraries worth the download
#

  • quantstats — tear-sheet analytics in one call.
  • skfolio and Riskfolio-Lib — modern portfolio optimization (the subject of my Black-Litterman walkthrough).
  • vectorbt — vectorized backtesting at DataFrame speed.
  • QLib — Microsoft’s quant research platform (data pipeline + factor ML).
  • yfinance — the community-maintained scraper for Yahoo endpoints (handles the UA/rate-limit dance for you).

10. The graveyard (dead, and what replaced it)
#

These are the bookmarks I’ve had to bury — kept here so you skip the same dead ends:

Dead What happened Use instead
Quantopian shut down 2020 QuantConnect, Numerai
Quandl absorbed by Nasdaq Nasdaq Data Link
IEX Cloud sunset 2024 Polygon, Tiingo
feeds.finance.yahoo.com RSS retired (404) Google News per-ticker feeds
Reuters public RSS closed ~2020 Benzinga, FinancialJuice
CNBC /rss/ index page moved per-feed URLs like the top-news feed
optionsclearing.com domain retired theocc.com
13F List domain gone HoldingsChannel
OpenCalais discontinued GDELT
old FBI NICS PDF links moved fbi.gov/cjis/nics
Futures Magazine defunct —

A 403 or 401 almost always means bot-gated, not dead — test in a browser before deleting a bookmark. A 404 with a redirect to a paid homepage, like the old Yahoo RSS, means the service itself moved on.

Tested 20 Nov 2026. If you find a dead link, the comments are open — this page is meant to be alive.

Related

Pricing Options with Monte Carlo in NumPy — Verified Against Black-Scholes

Black-Scholes gives you a closed-form price for a vanilla European option in one line — so why bother simulating? Because the moment an option’s payoff depends on the path (Asian, lookback, barrier) or you add non-linearity (American exercise, Vol surfaces, XVA), the formulas stop existing and Monte Carlo keeps working. In this post we build a complete Monte Carlo pricer from scratch in NumPy, verify it against Black-Scholes, then extend it to path-dependent payoffs and Greeks. Every number shown comes from a single run with a fixed seed (numpy.random.default_rng(42)) — you should be able to reproduce all of them.

AR(1) Process Autocorrelation: Theory vs Simulation in Python

You simulate an AR(1) process with parameter ρ = 0.7 and compute its sample autocorrelation at lag 1. What value do you expect? The theoretical answer is simple: the lag-1 autocorrelation of an AR(1) process is exactly ρ. But how quickly does the sample autocorrelation converge to this value — and how noisy is the estimator at realistic sample sizes? In this post, we derive the theoretical autocorrelation function of an AR(1) process, simulate it in NumPy, and verify both the point estimate and its dispersion against theory with a Monte-Carlo experiment.

Kalman Filtering from Scratch in Python: Linear State Estimation with Ground Truth Verification

You observe a system’s position every second with a noisy GPS sensor, but you need to know both position and velocity for control. You have a physics model (constant velocity) but the measurements are unreliable. How do you recover the true state? The Kalman Filter answers this question: it estimates a system’s hidden state from noisy measurements, combining a prediction from a dynamic model with a correction from new measurements, weighted by their respective uncertainties. In this post, we build a complete Kalman Filter from scratch in NumPy, verify it against a known ground-truth trajectory, and show exactly how much uncertainty it reduces compared to raw measurements.