Methodology — How we track trends

Source

Source: Google Trends trending searches for the United States, via its RSS feed (geo=US). This site is not Google Trends and is not affiliated with Google. A wider CSV deep scan (about 480 items, 4 times per day) may supplement the RSS list for archive completeness; the RSS list (about 10–20 items) stays primary.

The 15-minute snapshot

A cron job fetches the feed every 15 minutes and appends one immutable snapshot per fetch to a JSON file for that day. Each snapshot records: the timestamp in UTC and in US Eastern time, the full ranked list, the search-volume band exactly as reported by Trends (for example 200K+), each trend's start time, and its attached source headlines. Snapshots are append-only and kept indefinitely. A missed fetch is recorded as a gap in the snapshot file and is never backfilled with invented data.

What is computed from snapshots, per topic

• First seen / last seen: the first and most recent snapshot containing the topic, in US Eastern time.
• Duration: the time between first and last seen, or “still trending” while the topic is active.
• Peak rank and peak band: the best rank reached and the highest volume band reported during the topic's life.
• Rank path: the topic's rank at every snapshot, rendered as a position chart on its topic page.
• Comebacks: the count of separate runs. A topic that leaves the list and returns starts a new run on the same topic page.

Categories and pages

Topics are grouped into sports, entertainment, technology, politics, business, weather and health. Each day gets one permanent archive page, each recurring topic gets one permanent topic page, and each category gets one permanent category page. New developments update the existing story and topic pages in place; they do not create a new page per hour.

Limits, stated plainly

• Volume bands are bands, not exact search counts. This site never converts a band into a precise traffic figure.
• The RSS list is short (about 10–20 items). A trend outside that list may exist without appearing here until the wider scan picks it up.
• Trends data reflects what people searched for, not what is most important. A high rank is a measure of search interest, not of editorial importance.
• Coverage is US only, in English, in US Eastern time. A topic trending elsewhere is out of scope for this site unless it also trends in the US.