Source Collection
Topic Radar periodically fetches recent public post lists from selected communities. It reads titles, basic public list metadata, and source links. It does not require user accounts, private messages, or non-public community data.
Term Extraction
The system extracts candidate terms from recent titles, filters common filler words, and groups known aliases where practical. For Korean text, the local prototype uses a morphological analyzer during development; production rules may vary by host environment.
Ranking
The main board ranks terms by recent mention count. A word appears larger when it appears in more titles during the current scan window.
Rising Terms
A term is marked as rising only when it increases meaningfully against the previous snapshot. Small changes, including ordinary +1 movement, are intentionally ignored.
Limitations
Community titles are noisy. A term can trend because of jokes, spam, arguments, or repeated low-quality posts. Topic Radar should be treated as a signal board, not a factual verification tool.