Detect emerging, bridge-candidate, and undercited papers inside a curated audio-ML corpus, then expose the signals behind every recommendation.
Trends
Counts and deltas are computed over included worksin Research Radar (the same ingestion policy as search and ranking), not OpenAlex-wide. We compare papers from the selected "recent" year window against an earlier band to label topics rising, steady, or cooling: a sanity check for the corpus, not a field-wide forecast.
Topic labels are imported metadata and can be noisy; use them as coarse navigation hints, not authoritative classifications.
API returned 503 for /api/v1/trends/topics
If a label rises here, it means your current venue mix and inclusion rules are producing more papers tagged with that OpenAlex topic in the recent window: useful for interpreting topic_growth-style signals, not for claiming the whole field moved.
Corpus-snapshot-specific trends, cluster dynamics, and labeled evaluation (see Evaluation) sit outside this v0 view.