Concert ticket prices fluctuate in an unexpected manner before the actual event. Instead of overpaying for bad tickets for the rest of our lives, we aim to build a data-driven recommendation system that predicts the optimal day to buy a ticket for a specific concert or artist, using time series forecasting and machine learning regression within a serverless (AWS Lambda) pipeline.

A serverless machine learning pipeline that transforms real-time concert ticket data into actionable purchase recommendations using time series forecasting and predictive analytics.

Concert ticket prices fluctuate in an unexpected manner before the actual event. Instead of overpaying for bad tickets for the rest of our lives, we aim to build a data-driven recommendation system that predicts the optimal day to buy a ticket for a specific concert or artist.

This project focuses on predicting optimal purchase timing for concert tickets to help consumers make informed decisions and avoid overpaying.

• Predict optimal purchase timing for concert tickets across multiple artists and venues
• Deliver actionable insights through an interactive recommendation system
• Support data-driven decision making for concert-goers
• Benchmark multiple forecasting approaches using time series analysis and machine learning regression

Leveraging real-time concert ticket pricing data, we construct an automated serverless pipeline (AWS Lambda) with:

• Temporal features: Time-to-event, day-of-week, seasonal patterns
• Price dynamics: Historical trends, volatility measures, rolling statistics
• Event characteristics: Artist popularity, venue capacity, location data
• External factors: Demand indicators and market conditions

Models are evaluated using R^2, RMSE, MAE, and directional accuracy to assess predictive performance and real-world utility for purchase recommendations.

• Unit of Analysis: Geographic market × Year
• Time Period: 2022-2025 (4 years))
• Key Variables:
  • Outcomes: gross, tickets, shows, avg_price
  • Lagged: gross_lag1, tickets_lag1, etc.
  • Growth: gross_growth, ticket_growth
  • External: Ticketmaster event counts and prices

• Unit of Analysis: Artist × Year
• Time Period: 2025 only
• Key Variables:
  • Performance: num_events, avg_ticket_price, unique_venues
  • Spotify: artist_popularity, artist_followers
  • Derived: events_per_venue

Data Science Team (Alphabetically): Harini Mohan, Joshua Piña

Institution: Georgia State University

Report bugs or request features

• Joshua Piña: [email protected]
• Harini Mohan: [email protected]

👉 We Could be Prophets...