What people buy, where they go and what they share tells a story. Increasingly so with data.

In the case of public events, understanding crowd behavior and consumption can help make or break an event’s reputation where guests’ safety and comfort is at stake. Entertainment venues and events in particular generate a ton of data from a myriad of sources: attendee posts on social media, ticketing apps, Spotify listens and chip-based vendor payments.

How can such data be best collected, analyzed and ultimately used? In the summer of 2018, Denmark’s Skanderborg Music Festival (Smukfest) undertook a proof of concept in a live setting.

A team of IBM volunteers was plucked out of the confines of a normal office environment and spent eight days at the festival, working around the clock, working in a makeshift container and sleeping, when possible, in tents in the forest environment. The team’s goal was to capture and analyze data that would, in turn, help build a better experience for festival attendees, at times, having to iterate in real time.

Learn what happened when IBM brought Watson to Denmark’s most diverse summer festival during a special Think 2019 presentation (6656) on Wednesday, February 13 from 5:30 PM to 6:10 PM at AMC at Metreon in Theater 14.

An IBM team works around the clock on four use cases

Smukfest, held each summer in the woodlands of Skanderborg, hosts more than 150 world-renowned performers each year, such as Britney Spears, Shawn Mendes and deadmau5, plus up-and-coming Danish acts.

The festival’s bars, shops and installations are made from 100 percent recyclable materials, creating a whimsical atmosphere where attendees are encouraged to escape from the norm in a peaceful setting.

Managing a large festival in a rural forested environment for 60,000 attendees is a considerable logistical exercise to run smoothly, profitably and safely. In the era of smartphones, there is increasing pressure to offer stronger customer mobile engagement and respond to factors such as weather, crowd demand and supplies. Thanks to a sponsorship from IBM Denmark, the festival could get unique insights in real time from a Watson chatbot on food and beverage buying behavior, crowd safety, audience preferences and frequently asked questions.

The festival collects an unprecedented level of data from around the event from the use of rental vehicles to attendee location tracking to number of security teams.

IBM volunteers worked on four predetermined use cases based on available data, harnessing streaming analytics on IBM Cloud to prepare data, IBM Compose and Cloud Object Storage for storing the data, Watson Studio for analyzing the data and the IBM Message Hub for powering apps:

• Using transaction data from the purchase of food and beverages to help increase sales and decrease the time taken queuing
• Analysis of attendee location data to help reduce overcrowding at stages and other potential safety hazards
• Analysis of logistics patterns of vehicles used in the setup and cleanup operation to help identify areas for improved efficiencies
• Development of a chatbot to provide festival goers with practical information

Data collected while privacy was respected

Smukfest uses wristbands with Near Field Communication chips to protect the digital user experience, allowing guests to securely pay for tickets, VIP festival access, food, drinks and accommodations amounting to 745,495 cashless transactions.

Other data sources include included Spotify playlists,geolocation of guests and security and medical teams, plus bars, stalls and restaurants.

The IBM volunteers collected and analyzed the data, cleaning out all digital traces of personal information to comply with the EU General Data Protection Regulation (GDPR). After washing the data, what was left was all the transaction data, purchases, geolocation and Spotify data, cleansed of all personal information.

From here, the team followed every transaction from bars and restaurants, comparing categories with products over time and by revenue generated. This allowed the team to see how, for example, typically poorer tequila sales were tracking against the more popular gin and vodka drinks, which vendors were performing better than others, and even what time of day beer sold best.

Using a data lake to combine data sets

The team took festival goers’ geotracking data and mapped it in a timeline over the day, detecting patterns of movement around the festival area within the 600,000 square meters (148 acres) comprising the festival grounds, including parking and camping areas.

The team then combined this heatmap of guest movement with geotracking data on the medical and security teams to made quick decisions on deployment to more crowded areas of the festival.

“We had this huge amount of data in the forest but it was in silos. We didn’t have access to all the data and didn’t have any chance to combine data,” said Smukfest CIONiels Mørk.“We’re not just organizing and analyzing data, but combining different kinds of data sets to give us different points of view on topics that can support our decision making.”

Now that Smukfest organizers know it’s possible to gather all data sources and have them made available instantaneously, the use cases for 2019 will undoubtedly be more complex and challenging.

Can you guess what liquor had the best sales? Can you think of a great use case for Smukfest 2019? Attend a special Think 2019 presentation (6656) on Wednesday, February 13 from 5:30 PM to 6:10 PM at AMC at Metreon in Theater 14.