[Kamal Datta] Welcome, Marcel. Glad to have you on the pod. [Marcel Altenburg] Thank you. Thank you for the invite. [Kamal Datta] Great. So I'll start with something I read on a World Major Marathons article that was published recently, and then I think we'll introduce you after that. In 2024, the London Marathon broke its record with 51,000 runners, up from 43,000 runners in 2023. Five months later, the BMW Berlin Marathon celebrated its 50th anniversary with a record 54,280 runners. But that celebration was short-lived because, in November, the New York City Marathon set a new record with 56,646 runners. Now, all of these races—London, Berlin, and New York—had one man in common who helped make it possible to increase the size of the field, and that is you, Marcel Altenburg. You are the senior lecturer in crowd science at Manchester Metropolitan University in Manchester, UK. So, welcome, Marcel. What did I miss? [Marcel Altenburg] Thank you! Yeah, that’s how they write it. But I need to also say I work with great teams at each of these races who put all of this into operation. Here in Manchester, we focus on crowd science, working closely with these events to improve them—making them safer and allowing more runners to participate. We provide the data for race organizers in London, Berlin, New York, and let’s not forget Chicago—they had a fantastic race last year with a world record. Just had to mention that! These teams take our research and recommendations and turn them into real-world improvements for their events. And last year, we had a blast. It was an intense year, but now looking back, it was worth it—we had the biggest and best races ever, allowing the most runners to enjoy their experiences. [Kamal Datta] Cool. Now, for an average Joe like me—you know, I’ve run four of the six majors, and I know the seventh is already there, and I need to plan to eventually run the eighth and ninth. What exactly is crowd science, Marcel? Just break it down in simple terms. [Marcel Altenburg] Let’s not start from my perspective; let’s start from yours. I know you ran Berlin last year, and a lot of things happened there. You were part of a crowd, with your own race plan, experiences, struggles, and victories. What we do is collect all those stories, all that data, and study it. Then, we use it to predict and anticipate future races. Take Berlin’s 50th anniversary, for example. They wanted to introduce new features, and we helped them predict what those changes would look like—from movement patterns to crowd psychology, all while ensuring safety and optimizing the race experience. Crowd science applies to marathons, but it also extends to festivals, airports, and transportation hubs. The core principle remains the same: understanding how people move in large numbers. [Kamal Datta] Wow. So you basically help design the race experience, from the start line to the finish line, for organizers to be better prepared? [Marcel Altenburg] Exactly! I like how you said that. From start to finish, that’s the whole race—26.2 miles, crossing boroughs, cities, landmarks. Race organizers want the event to be perfect, but they also need to keep it safe. The start itself can last anywhere from one and a half hours to three hours, depending on the event. It’s crucial to ensure every runner gets a great experience while maintaining safety. Now, imagine you wanted to organize your own race. You’d have many plans but also many questions: How do I handle thousands of people? What happens at aid stations? How many tables for water do I need? Will toilets get overcrowded? Those are the same concerns organizers face. It’s essentially like planning a massive birthday party—but instead of guests, you have tens of thousands of runners. That’s where crowd science comes in. [Kamal Datta] Interesting. Let’s take the example of the start line—New York City Marathon has one of the most complex start setups. With three starting points and their own unique challenges, how do you even begin distributing runners? And how does the system determine that I, Kamal, should be assigned to Corral G at a specific time? [Marcel Altenburg] Great question. Think of it like a recipe. The start is the last point where we interact with runners before they begin their race. Everything up to that moment is planned meticulously to ensure a smooth experience. On your bib number, you’re given information about your wave and corral. Your wave determines when you start, while your corral determines where exactly you should be within that wave. It’s similar to boarding at an airport—you might be at the same gate, but different boarding groups and seat assignments dictate your placement. The timing of wave launches isn’t always evenly spaced—it could be 7 minutes and 20 seconds between waves, or 10 minutes, depending on the plan. Then comes ordering within waves. If you’re a faster runner than me, starting behind me would mean overtaking me. That affects the flow of the race. If I start behind you, there’s no overtake—it’s a small but crucial difference that shapes the race experience. Space matters too. If the start road is narrow, runners are positioned behind each other. If it’s wider, they may be placed side by side. All of this goes into the algorithm, which calculates the impact of these decisions. [Kamal Datta] Wow. So the algorithm starts by looking at the finish line experience first? [Marcel Altenburg] Yes! We plan backward from the finish to the start. The finish area has physical constraints, and race organizers know their city inside out. We incorporate their measurements into our algorithm, factoring in pinch points, aid stations, and any logistical challenges. Additionally, we aim for a steady flow. We don’t want runners to come in waves, causing congestion at the finish line. Ideally, we maintain a predictable density, ensuring every finisher has a smooth post-race experience. [Kamal Datta] Fascinating. I read that you predicted the peak finisher flow at the New York City Marathon down to a single runner—1,366 predicted, and the actual number was 1,367. How do you get so close? [Marcel Altenburg] It’s a combination of solid science, years of experience, and massive amounts of data. The more data we have, the more precise we can be. Runners produce millions of data points, and each year, our algorithm gets stronger. At New York, we predicted peak flow so organizers could prepare volunteers, medics, and logistics accordingly. Being off by one runner in a five-minute window is remarkable accuracy. [Kamal Datta] Incredible. We’ve covered planning, real-time adjustments, space constraints, and weather considerations. Any final thoughts on how crowd science benefits events beyond running? [Marcel Altenburg] Absolutely. We’ve applied the same principles to concerts, festivals, airports, and even major global events like the Dubai Expo, the Olympics, and football championships. At non-running events, data is harder to come by—ticket scans might be the only input. But our algorithm helps complete the picture, providing insights on crowd behavior, flow, and optimizing placement of services. In the end, we help make large-scale events safer, more efficient, and more enjoyable for everyone. [Kamal Datta] Marcel, this has been a fantastic conversation. Thank you for sharing your expertise with us. [Marcel Altenburg] Thank you, Kamal! And to all your listeners—check your bib number. There's a plan behind it, and it's designed to give you the best race experience possible.