Thanks to Brilliant for supporting this episode of SciShow.

Go to Brilliant.org/SciShow to learn how you can take your STEM skills to the next level with their new course on geometry fundamentals! [♪ INTRO]. Five hours of off-road driving from the city of Kassala, in Eastern Sudan, lies a collection of ancient tombs.

Researchers have written about them for years, but they aren’t as well-studied as other archaeological sites because of their remote location and their number. There are a ton of tombs. Enough that examining them one by one would take too much time.

So archaeologists resorted to a less traditional way of studying them. They used a statistical method designed to study something a little different...the stars. There are two main types of tombs in this area: there are tumuli, which are simple raised structures made of earth and stone.

And there are qubbas, square shrines, or tombs constructed with flat slabs of metamorphic rock. These tombs come from two different periods. Qubbas are from the medieval Islamic period in the area, which ended sometime in the 15th century, while the tumulis date back even further to prehistoric Africa.

This overlap of cultures makes these tombs of particular interest to archaeologists, because these tombs can contain information about the cultural landscape and history of the region. And we still don’t know how many there are, but we know that they number in the thousands. There are so many that counting and cataloging them all by hand would require multiple trips out to this remote region and a lot of work.

In the past, archaeologists might not have had much choice but to plan an expedition and look at each tomb by hand. But nowadays, modern archaeologists can use technology to make the task a lot easier. And in the case of the tombs in Sudan, archaeologists realized that they could use a statistical tool to model how the tombs were arranged, which had previously been developed to model stars.

Satellites have been used in archaeology before to get an eagle-eye view of remote regions where exploration could be challenging. And the site in Sudan, well it would be a perfect candidate for satellite work. But while satellites can take a nice aerial picture of the site, they aren’t like smart or anything.

They can’t exactly determine which tombs are related to each other. And that’s an important aspect for researchers studying funerary practices and traditions. So for that, a different tool needed to be used.

The Neyman-Scott Cluster Process. Scientists developed it to understand how stars and galaxies in the sky are grouped together. It’s a type of cluster analysis.

Which is a way to tell how random points can be grouped together and how they relate. So, for example, if you’re doing market research, a cluster analysis might tell you that customers share something, like an age, or a zip code, or some other demographic, just by looking at each individual’s behavior. You could find that one cluster of customers are mostly millennials, and you might find another that’s mostly Gen Z.

But when you turn cluster analysis to astronomy instead of economics, these types of statistical analyses can look at stars as points on a map and tell you which stars are likely to be grouped together. The thing is, normal cluster analyses look at each point as an individual, independent from all other points. And that isn’t exactly how stars work.

Stars and galaxies are affected by gravitational attraction, which means that they tend to be associated more with certain stars than others and might even group around a central point. So to properly track the movement of stars, you need a system that considers points that might be connected to each other. And that is where the Neyman-Scott Cluster Process comes in.

Instead of just looking at individual points, this process detects things that are called “offspring points,” which tend to cluster around “parent points.” That means it looks for points that branch out from each other. For example, regular cluster analysis might help you find clusters of trees when looking at a rainforest. But the Neyman-Scott process could help you find regularly-spaced parent trees, with their offspring trees clustered around them.

It’s the fact that this process takes relationships into account that makes it useful for archaeology. Satellite images of the Kassala site show that the arrangement of the tombs depends on the topography of the area, which makes sense. Because tombs were more likely to be built in foothills or on top of ridges than in slopes.

That’s because it’s harder to build on a slope than it is to build on something flat. But when the researchers applied the clustering process to these images, they found six main clusters of qubbas, and not all of them could be explained by topography. Using the Neyman Scott Cluster Process, the researchers found that there seemed to be some kind of social gravitational attraction.

Just like with the stars, qubbas were clustered around central points, possibly connected to tribes and families. These qubbas also tended to cluster around the much older tumuli for reasons that are yet unknown. By comparing the arrangement of the tombs with historical sources, researchers were able to find that these qubbas were probably tribal or family cemeteries of the Beja people, who lived in the Kassala region.

One interesting thing, the qubbas in Kassala are located 60 kilometers south of the site previously thought to be the southernmost occurrence of Islamic Beja funerary monuments. Which means that scholars underestimated the number of qubbas in the region and their location. The Beja people have settled in the region since time immemorial, and ancient written records support them being in the region before 2000 years ago.

So the qubbas and tumuli might be more culturally connected than what researchers previously thought. Of course, the authors could have found all this out the “old-fashioned way” by exploring the region and looking at each cluster separately, but this would have taken much more time and much more money. But now, researchers can not only use statistics to unlock the secrets of the universe, they can also apply those same techniques to unlocking the secrets of our history.

If you want to up your game on statistics, you should check out today’s sponsor, Brilliant. Brilliant is a website and app built off of the principle of active problem solving: you learn best while doing and solving in real-time. And they’ve overhauled their courses to be even more interactive.

One of their newly updated courses is geometry fundamentals which you can use to build a strong foundation in statistics! With their course, you can discover how intuitive geometry can be when keeping assumptions simple while using your logic and reasoning to set up calculations. So, if you would like to check them out, visit brilliant.org/scishow, where you can get 20% off an annual premium subscription to Brilliant.

And thank you for your support! [♪ OUTRO].