THE BIG DATA LAB AT ROSKILDE FESTIVAL – TOILETS, BEER and CONDOMS

THE BIG DATA LAB AT ROSKILDE FESTIVAL – TOILETS, BEER and CONDOMS

At this years‘ edition of ROSKILDE FESTIVAL (RF), the well known outstanding music and art festival in Denmark, we found an IT based issue to talk about: BIG DATA. entertaim.net’s Martin Hannig  had the  opportinity to visit the IT Big Data center at RF right before the festival start ….

by M. Hannig, pic by cbsBDA,

BIG DATA: Nowadays this term is in everyone’s mouth, and economy and society should take notice on Big Data. Big Data means: collecting data as much as possible, combine it while using algorithms on it, and thus make precise predictions about the future’s reality. It’s more than market research, because Big Data is flowing in – you don’t need to ask somebody, moreover: people actually just don’t take notice that data are being collected.  As complete industries and services are changing because of Big Data usage, it’s just a matter of time while the live music industry will take part. Per Østergaard Jacobsen from Copenhagen Business School (CBS) is the first to research on this, and his playing field is RF, where he and his team were trying to apply their research to festival reality. Per has a personal background that connects him to the festival industry: he was a tour manager before he started working at University, and he worked 3 years as a strategic consultant for RF – so he knows the festival game and he surely knows RF.

The project started back 4 years ago when Per had the idea to use RF as a lab for implementing sustainability issues by using IT. The transformation into practice took a few years; Per told in the beginning they were very naive. They worked a lot with Excel sheets and documents. When they started to realize that there is a lot of live data around they found out that Excel isn’t the right tool. So they ended up with some business partners like IBM, to establish the 1st lab in 2015. Main goal 2015 was to find out about the data quantity and quality – and they found out a lot.

In 2016 there are 45 colleagues  from Denmark, Netherlands, Finland, France, Germany and USA working in the Big Data lab to reach the next step: to find out how to use this data coming live from the festival ground. Still under the influence of sustainability aspects, but also to create mutual value for both customers and companies. Per and his team try to be a learning lab that brings ideas out to the real world and of course to the festival. He mentioned that it’s an extraordinary situation to have the data of a “city” of 130.000 people – the whole festival delivers data and the lab tries to make something useful to give it back to the people. It’s a matter of learning by doing – maybe this year they find out new data patterns to work on in 2017. And it’s clear: the whole big data lab wouldn’t be possible in a commercial festival surrounding. It helps a lot that RF is a nonprofit organization – so there’s no profit aim behind this research.

Per explained the heart of his project, standing in front of a big screen showing live data on the wall. It’s all based on the Roskilde Festival app – it delivers GPS driven data to the data cloud, so in the IT data center we are able to see the concentration of guests in the different festival areas. The grade of concentration is shown by different colours: e.g. red means high concentration. At the beginning of the festival we could watch big red areas at the gates – and they were moving to the different stages minutes after opening of the gates. Per’s team uses this geo located data to find out where the people are at the moment. If you know the number of fans going to Orange Stage right now, you can optimize service and of course security. On the other hand you can plan the beverage delivery of the stalls. Another practice example of using data: if you have 60.000 fans in front of Orange Stage und you combine this information with the weather forecast, you might find out that it will be raining in 60 minutes – and you have to bring out like sand on some weak ground to avoid muddy situations. In the end the festival could optimize the service comfort for the guests. The team even handles data by tracking each single person, e.g. how far they walked, who is walking in the morning, who in the evening, in the night, and where they are going – and they want to be able to predict that behavior. The main aspect of this person tracking is the health prospect: to see how many calouries are burned by people during the festival. Many of the festival guests are not active on a daily basis; most of them are not runners or playing football. As they are quite active during the festival days just by walking so much kilometers, the team hopes to get them being aware of these activities, and nudge them continuing the exercises back home after the festival. Nice idea and a wonderful interpretation on using Big Data for health reason!

But there is another example of using big data at Roskilde Festival. It’s based on the internet of things. We all know it: all kind of things could be connected to the internet – from cars to wearables. Per’s team is going further: they connect toilets to the net! A joke? No, it’s a quite serious effort to rise the comfort for the guests: the “toilet tracker”. In a few weeks‘ work, the team planned the best sensor measurement points and then prepared some toilets with 4 sensors inside the toilet to collect data! Data is then being send by wifi to a special gateway and then into the cloud. People using the toilet don’t know about this experiment – it’s anonymous data, and for good reason the team wants to have normal behavior…

Engineer Jan Ekström from IBM Denmark  – one of the supporting companies  – shows a monitor with toilet real time data made out of different lines and numbers.  One number is showing the current persons weight, another the amount of “material” in the toilet. Also one can read the cumulated numbers of material in one toilet during the last days.  It’s also visible how many persons have used the toilet. The practical use is clear: one can foresee and predict when to clean up the toilets – classic Big Data analysis.  Or in which intervals of time it should be cleaned – the festival managers may optimize it and avoid waste of time and energy for unnecessary cleanings. They even may optimize the several rounds by the cleaning vans – which sums up to the RF sustainability efforts. Or the amount of paper and soap to deliver. Or the number of toilets which are overloaded and must be cleaned up urgently – what’s more horrible on a festival than to be forced to use a full filled toilet? In the end, the whole toilet project provides more information than Jens thought it could. Moreover: Jens told me about a possible practical use for airlines – as every gram of things in a plane costs pretty much Kerosene it could be interesting to reduce the average amount of human waste in plane toilets.

Another quite surprising big data analysis the team made on the correlation between beer sales and condom sales (no fun!). PhD fellow Niels Buus Lassen from CBS showed it on another monitor: at the first festival day on the campground, Saturday, he found out the peak of beer sales at 8 p.m. – and 2 hours after that there was condom sales peaking! Random? There’s exactly the same pattern on Sunday. Same on the following days. There’s a clear correlation between beer and condoms sales – later peaking one to two hours after the beer sales.  Funny enough condom sales go to zero at 7. p.m. – dinner time! And after dinner the condom sales takes off again. Niels also tried to find out about beer and headache pills correlation – but he found no patterns on this…