ClinicalTrials.gov sponsor pipeline velocity (monthly)

Monthly enrichment signals scoring clinical-trial pipeline momentum for 49 curated pharma / biotech / medtech sponsors (USA, Canada, Europe, Japan, Israel). Each month the keyless ClinicalTrials.gov v2 API (U.S. National Library of Medicine) is queried per sponsor for study counts (countTotal only — no study records are downloaded) with StartDate in the trailing 12 complete months vs the prior 12 months, each for all phases and for Phase 3. Per sponsor: trial-start counts per window, trailing-vs-prior velocity_pct ((cur - prev) / prev * 100; null when the prior window is empty), the Phase 3 analogue, pipeline_score = 100 * (50% min-max(log1p(starts)) + 50% min-max(velocity clipped to [-100%, +300%] mapped to [0, 1])), min-maxed within the snapshot so it is a relative, not absolute, momentum gauge, deterministic pipeline_rank (ties by sponsor name), and a momentum_flag (surging / growing / stable / declining / contracting / new_entrant / dormant). Sponsors with no starts get an explicit dormant row — the panel is join-complete. Who joins this: a pharma / biotech BD team joins trailing-12m trial-start velocity per sponsor to their licensing-target watchlist on sponsor_name_normalized + as_of; a clinical-operations team joins Phase 3 momentum on sponsor_name_normalized. Primary key: (as_of, sponsor_name_normalized); join keys: sponsor_name_normalized, sponsor_name, as_of, country_code (ISO alpha-3, curated HQ). Caveats: query.spons uses tokenized matching over the NLM-indexed sponsor / collaborator name fields, so a count can include studies where the sponsor is a collaborator; studies with no StartDate are invisible to the date filter; the registry is sponsor-submitted and backfilled, so a same-month re-ingest can shift counts slightly; pipeline_score is within-snapshot relative and not comparable across months. No study text is stored. Sample use: filter momentum_flag = 'surging' for sponsors accelerating their pipeline, or order by pipeline_rank for this month's strongest trial-start momentum.