Overview
VenueMatch is a deployed booking-intelligence platform designed to help independent music teams evaluate where an artist is most likely to fit. It converts fragmented artist, venue, market, and event data into an explainable recommendation rather than a black-box score.
Problem
Artist routing and venue booking decisions depend on several signals at once: genre compatibility, local demand, booking history, room capacity, and artist pull. Reviewing those signals manually is slow and makes it difficult to compare options consistently.
Dataset / Inputs
- Live records covering 336 artists
- 52 venue profiles
- 428 artist-event relationships
- Genre, city, historical booking, capacity, and popularity signals
System Architecture
- Structured ingestion layer for artist, venue, and event data
- Deterministic database upserts for repeatable refreshes
- Explainable scoring engine with weighted recommendation factors
- Protected Vercel cron jobs for daily ingestion
- Responsive web interface for artist-to-venue, venue-to-artist, and market exploration workflows
Methodology
The recommendation engine combines genre fit, city demand, venue history, capacity compatibility, and popularity. Each factor contributes to a visible score breakdown so users can understand why a recommendation ranks highly and where the fit is weaker.
What I Built
- Deployed the end-to-end booking-intelligence application
- Designed the explainable recommendation model and score breakdown
- Connected live artist, venue, and event relationship data
- Automated daily ingestion through protected scheduled jobs
- Implemented deterministic upserts to keep refreshes stable and repeatable
Results / Proof Points
- 336 artists represented in the live system
- 52 venues available for matching
- 428 artist-event relationships supporting recommendation context
- Daily automated ingestion and deterministic database updates
- Publicly deployed interface with multiple recommendation workflows
Business Value
VenueMatch gives booking teams a consistent way to compare artist-venue fit, inspect the reasoning behind each result, and focus research on the strongest opportunities instead of manually reconciling disconnected signals.
Limitations
The current scoring model is a decision-support system rather than a causal predictor of ticket sales. Recommendation quality depends on the coverage and freshness of the underlying market and booking data.
Next Steps
- Add richer market-demand and historical performance signals
- Evaluate recommendation quality against observed booking outcomes
- Improve confidence indicators for sparse artist or venue histories
- Expand evidence links as technical documentation and project reports are completed