How the radar works
The radar watches open, permitted data sources for each niche: Wikipedia pageviews (Wikimedia), news coverage (the GDELT Project), new research papers (arXiv and PubMed), Hacker News stories, podcast episodes (Podcast Index) and headlines from curated niche publications.
Every day the pipeline counts how much attention and coverage each topic gets in each source, compares the last few days with the topic's own four-week baseline, and turns that into a trend score, a lifecycle stage and an opportunity score.
- Update frequency: Wikipedia daily (once the previous day's data is published), news and niche headlines hourly, papers and podcasts daily.
- History: public pages show the last 30 days. Full history will be part of Pro.
- No social platforms: we never use data from YouTube, Instagram, TikTok, X, Reddit or Twitch, and we never scrape. Topics matter, not platforms, so you can post about them anywhere.
- Headlines only: for news and articles we store and show only headlines, the publisher, the date and a link. We never republish article text.
If one source is unavailable, scores are calculated from the remaining sources and each topic page shows which sources were used. The full method is on the methodology page.
Didn't answer your question? Contact us.