Multi-signal momentum radar with confidence and coverage states.
Artist Decision Intelligence
A multi-provider decision intelligence system for turning fragmented music signals into market-aware, confidence-weighted artist and track insights.
Designed the product around signal provenance, provider coverage, confidence gates, and explicit sparse-data states so the system can separate useful evidence from weak or misleading signals.
01 / PROBLEM
What the system is trying to solve.
Music data is distributed across providers, markets, and incompatible ranking systems. A raw chart position or one-provider spike can look meaningful without enough evidence to support a decision.
02 / SYSTEM
How it works.
- 01
Collect and normalize observations from multiple music-data providers.
- 02
Group evidence into behavior families instead of treating every source as an independent vote.
- 03
Score momentum with provider caps, minimum-family requirements, and market-aware coverage checks.
- 04
Gate outputs when evidence is sparse, unavailable, or policy-restricted rather than forcing a recommendation.
- 05
Preserve provenance and audit results so every surfaced insight can be traced back to its evidence.
03 / ENGINEERING
Decisions that shaped the project.
Market-level analysis across multiple cities instead of a single global score.
Immutable observation windows and explicit leakage safeguards for longitudinal comparisons.
Provider-policy review gates, provenance tracking, and contamination checks.
Recommendation states that distinguish REAL evidence from DRAFT, SPARSE, ZERO_SIGNAL, and UNAVAILABLE conditions.
04 / EVIDENCE
A few concrete signals.
18
Candidate tracks4
Markets504
Family audit results0 · gated by evidence
Public REAL recommendations