Strategic Read this week ⚡ AI-WRITTEN
Generating…
Synthesizing the week's signals across panels.
Catalog at a glance
Format split
| Format | Videos |
|---|
Outlier score
⚡ AI insight
Reading the outlier panel…
Score = video views ÷ median of 20 prior uploads. Excludes the first 14 days post-publish, so early-life noise doesn’t poison the ranking.
🚀 Punching above their weight
Loading…
⚠️ Underperformers
Loading…
Cross-platform divergence (YouTube vs Spotify, residuals from log-log fit)
⚡ AI insight
Reading the cross-platform divergence panel…
🎧 Spotify-strong, YouTube-weak (audio audience exists, push to YouTube)
Loading…
📺 YouTube-strong, Spotify-weak (visual production carrying it, consider Spotify push)
Loading…
Cross-platform combined-rank (YouTube views × Spotify streams, attention-weighted)
Cohort distributions
⚡ AI insight
Reading the cohort distributions panel…
Distribution of latest views within each cohort. Bar shows min → max, box shows P25 → P75, vertical line marks the median. Sorted by median descending. Small-sample cohorts flagged.
By content type
Loading…
By featured-artist relation
Loading…
By content × relation
Loading…
Cadence × performance
⚡ AI insight
Reading the cadence panel…
Monthly upload cadence (bars) against per-video median views (line). Default view: last 36 months. Toggle to see the full catalog history.
Loading…
Not built yet
What this dashboard will show once Phase 2 lands
Outlier scoring, cohort distributions, cross-platform divergence, cadence, and the auto-generated Strategic Read are now live. Still pending:
- Weekly picks across the cover/original × tent-pole/mini matrix — needs candidate-song discovery flow + live ranking model
- Breakthrough Signals — videos reaching new audiences, distinguished from concerning drift — needs YouTube comment sentiment classifier
- Hindsight — 6-month-ago picks vs what was done vs what happened — populates after first 6 months of recommendations
- Recommendation cards with explainable evidence — per-pick feature-based + example-based rationale (Afchar et al.)