Every artist on the bourse gets a score from 0 to 1000, recomputed from live public data each time they are scraped. It answers two questions: how big is this artist now, and which way are they heading.
Not on the list? Type any artist into the search on the bourse: if they are not tracked yet, Magnate finds them on Deezer and Apple Music, scrapes all six platforms live and adds them to the index. From then on they are refreshed like everyone else.
How many people the artist reaches. Each audience count is log-scaled against a ceiling, so going from 10K to 100K counts as much as going from 1M to 10M.
How hard the audience listens, not just how big it is.
Whether the artist is working: releasing, touring, posting.
How many of the six platforms the artist could be verified on. It also tells you how much data backs the score.
log_scale(x, cap) = ln(1 + x) / ln(1 + cap), clamped to [0, 1] base = 1000 · Σ wᵍ·groupᵍ / Σ wᵍ (over groups that have data)
momentum = clamp(1 + spike + trend + acceleration + breadth + peak − volatility, 0.80, 1.30)
A sudden jump in listeners (a viral clip, a TV slot, a surprise drop) is measured against the artist's own normal day-to-day movement. +2% means nothing for an artist who swings 3% a day, and a lot for one who usually moves 0.2%.
cap · tanh(z / 2), z = Δ24h / max(0.5%, σ)A least-squares line through the log of daily audience closes. A steady climb with a clean fit (high R²) counts in full; a noisy series that happens to end higher counts for half.
cap · tanh(g₂₀ / 10%) · (0.5 + 0.5·R²)Is growth speeding up or fading? The last week's slope, projected over 20 days, is compared with the full-window slope.
cap · tanh((g₇→₂₀ − g₂₀) / 5%)Real momentum shows up everywhere at once. When Spotify, Deezer, Last.fm, YouTube and Bandsintown all move the same way, the signal is worth more than a single-platform blip.
cap · (up − down) / nAn artist at their all-time high is in a different place from one at a third of it. Full credit at the peak, neutral at 60% of it, the full penalty at 20% or below.
cap · clamp((L / peak − 60%) / 40%, −1, 1)Erratic audiences are riskier. Above 3% average daily swing, every extra point of volatility costs 1.5 points of multiplier.
−min(cap, 1.5 · max(0, σ − 3%))Signals start neutral (0) and switch on as history builds up. The 24h spike uses our own reading from about a day ago. For the ~590 biggest Spotify artists, kworb.net's daily listener table (also scraped through Context.dev, every 6 hours) supplies the 24h change and the all-time peak from day one. Trend and volatility need at least three daily readings spread over three days or more. Breadth needs two platforms with a 7-day comparison. The primary series is Spotify monthly listeners, or the platform with the longest history if Spotify is missing. tanh keeps every factor smooth and bounded, so one freak reading cannot break the scale.
| Reach | 85/100 | × 40% |
| Engagement | 91/100 | × 25% |
| Activity | 37/100 | × 20% |
| Presence | 5/6 platforms | × 15% |
| Base score | 769 | |
| 24h listener spike | +2.7% | |
| 20-day trend | warming up | — |
| Acceleration | warming up | — |
| Platform breadth | warming up | — |
| Peak proximity | +2.9% | |
| Volatility | warming up | — |
| Momentum | ×1.056 | |
| Magnate Index | 812/1000 | |
Momentum from 1 days of daily readings.
It describes public audience data. It is not a valuation, a forecast or a promise of returns, and it is not investment advice.
Platform counters can be wrong, delayed or gamed. If a platform can't be verified for an artist, it is left out, never guessed.
Momentum needs history. A newly discovered artist starts at ×1.000 and builds a full 20-day picture as daily readings come in.