Predictive Modeling Around Music Festivals

Does data make sense at a festival?


Music festivals are - for the most part - a wild, angry, debauched mess. From the outside, it looks like the ultimate challenge for data scientists - trying to keep track of a sweaty farrago of bodies intent on squandering themselves in every direction. Attempts to determine attendees’ likes and dislikes when many are not in their usual frame of mind may seem futile, but with 32 million in the US attending at least one festival a year, data analytics is necessary to maximize the vast revenues that can be made.

In 2014, the five biggest festivals combined grossed over $183 million in ticket sales, and that’s before taking into account sponsorships, merchandise, and food and alcohol sales. Coachella grossed a record-breaking $78 million over two weekends in 2014, and broke its own record in 2015 by taking in $84 million. Organizers are now looking at data to figure out exactly what festival goers did over the course of their stay and which elements of the event were most successful by using a host of sensors and apps. Using the insights leveraged, organizers can make improvements both in real time and ready for future events.

One of the ways that they are using data is determining which artists to book next year, by tracking how many people went to an artist’s performance and how many stayed following the previous act by putting sensors at entrances to tents. Jeff Cuellar, the vice president of strategic partnerships at AC Entertainment, works for the organizer behind four-day Tennessee festival, Bonnaroo. He notes that they use a non-traditional census to gather data points about festival goers, asking them psychographic questions like whether they put tabasco on tacos to build a picture of their personalities. This can be used to better predict the attendees at future events, so they can customize the line-up accordingly, introduce other attractions, and tailor their marketing efforts.

Festival organizers are not just using analytics to drive revenue, they are also using it to deal with the range of logistical issues that arise when trying to deal with hundreds of thousands of people in various states of sobriety. There were 450,000 at Coachella, and obviously, this presents a safety problem. They all also need to be fed and watered, and all the waste produced by this disposed off. With 130,000 attendees, the 2015 Roskilde Festival in Denmark is the largest culture and music festival in Northern Europe. At Rosokile, Copenhagen Business School used IBM Watson analytics in a cloud environment to analyze the behavior of attendees, producing a massive 91 million rows of data for every minute. In an interview with theCube, Cheri Bergeron, leader of client engagement and advocacy at IBM, explained that IBM had helped the business school set up a real-time data lab to produce heat maps produced by a mobile phone app that helped track the movements of their crowds, so they could get an idea of when people were most likely to go to food stalls, and where crowd build up may present safety issues. They looked at this alongside other factors such as weather conditions and artist affinity, and organizers were able to organize stalls and staff appropriately, as well as determine set times.

Many believe that music festival market is at saturation point. There are hundreds of new festivals popping up every year across the US, and millennials fork out a fortune for the privilege of attending. Music will always be the focal point of a festival, and the experience that people have is something that’s really unquantifiable. Data analytics should, however, help prevent the festival bubble from popping, for the time being at least.

University lecture small

Read next:

How Are Higher Education Institutions Using Analytics?