News Monitor: Only New Articles From Saved Watches
Pricing
from $2.10 / 1,000 articles
Saved watches over the GDELT Project's 15-minute news files: keywords, organizations, people, GDELT themes and news sites. Each run returns only the articles it did not deliver before, one row per article, with a free summary per watch.
Up to 50 watches, each an object with a type. {"type": "keyword", "phrases": ["battery recall"]} finds the phrases as whole words in the title (add "match": "all" to need every phrase). {"type": "organization", "names": ["Federal Reserve"]} and {"type": "person", "names": ["Christine Lagarde"]} look in the organizations and persons GDELT found in the article (add "matchTitle": true to an organization watch to look in the title too). {"type": "theme", "themes": ["ECON_BANKRUPTCY"]} takes GDELT theme codes, and {"type": "source", "domains": ["bbc.co.uk"]} news domains. Every watch takes a "label" and "excludeTerms" (words that drop an article when they are in its title). The first run delivers the newest articles of the look-back as a baseline; later runs only what is new.
How far back a watch's first run reads, from 1 to 72 hours. GDELT publishes a file every 15 minutes, so 24 hours is 96 files. Later runs read only the files published since the last run, never more than 72 hours back.
The most articles one watch delivers in one run, newest first. A watch's first run is its baseline: its newest articles up to this number. When a later run finds more new articles than this, the newest are delivered, the older ones are left out for good, and the watch's free summary row says how many.
Also read GDELT's translingual files: news in other languages, machine-translated by GDELT. Names and themes come from the translation; titles stay in the original language, so keyword watches match those titles as written. These files are larger and are published about 15 minutes after the English ones.
Optional. Only articles from these news domains (subdomains match too), such as bbc.co.uk.
Optional. Articles from these news domains are left out.
Optional. Only articles that mention a place in one of these countries, as GDELT found them: a 2-letter FIPS 10-4 code as GDELT writes it (US, UK, GM for Germany, CH for China, JA for Japan) or a country name.
Optional. Only articles with GDELT's tone at or above this, from -100 (most negative) to +100. Most news sits between -10 and +10.
Optional. Only articles with GDELT's tone at or below this, such as -2 for clearly negative news.
Optional. The name of the key-value store that remembers what each watch delivered. With a name, such as my-brand-watch, an added watch delivers only its own baseline. Left empty, a run of a saved task keeps one memory per task, which works the same way; a run outside a task names the memory after its watches, so adding or changing a watch there starts a new memory and the next run delivers every watch's baseline again. Letters, digits and dashes.
