Geo Big or Go Home: Social Geolocation Data From the Winter Games

When you think about social data, what’s the first thing that comes to mind? Posts, Tweets, pictures, videos? Sure - analyzing the content of social conversations can generate great insights. But today I'm going to dive into a social data opportunity that gets the cold shoulder all too often.

Today we’re going to talk about where you at, son.

That’s right, today’s study is about three aspects of social data: location, location, and location. Specifically, geolocated social media posts and the insights they can lead to. And what better occasion to do so than the meeting of nations: the Sochi Winter Games?

First, let’s talk about the data. When location services are enabled by a user, many posts and Tweets will include precise coordinates in addition to the content of the post. For every post that has geolocation enabled, you’ll see their latitude, longitude, and location name from where the post was uploaded. What can we do with this data? We’re about to find out.

Before we make our way to the Winter Games - here's some details about the data set I created for this study:

  • All data in this post was collected via public APIs (yes, you can do this yourself at home.)
  • For Instagram I analyzed data from 80+ Winter Game athlete’s accounts via API, and then identified and grabbed posts that included location information during the games.
  • On Twitter, I used the Twitter API to check the profiles of 500 athletes to see how many share their location in Tweets. Luckily for this analysis, a good number of them (around 30%) did just that.
  • All in all, we’re looking at more than 1,400 geolocated social media posts since the start of the Games, just from the athletes themselves.

Let’s see what location data can tell us about the the Winter Games, how the athletes spend their time in Sochi, and Russian swimming pools in February. Here we go.

Winter Games Athletes: World Travelers

The Winter Games doesn’t just land in your lap - athletes travel across the globe competing to qualify to represent their countries. Let’s begin our exploration of geolocation data by mapping one athlete’s journey to the Winter Games, all based on data provided by his social media posts.

So where in the world is Mitchell Malyk, Canadian Luger? Or should I say: where has he been? We’re about to find out.

Looks like he has been racking up some frequent flyer miles. As you can see in the past few months, he’s been throughout Germany, Latvia, and finally arriving in Sochi with his teammates. And we can follow his progress based solely on his Twitter geolocation data.

But we can get even more detailed with this latitude/longitude data. With the above view, we’re tracking Mitchell’s travels at a high level by identifying his current country and city- in a bit you’ll see how detailed this data can be. From my tests posting geolocated media throughout a few different corners of the Mass Relevance office, this data can be accurate within around five feet. Five feet, people.

Onward to Sochi

Now let’s move our lens to the Winter Games. In aggregate, location information can show us patterns of usage and activity across a wide geographic region. Here’s a map of Sochi, with all geolocated posts from athletes during the Games. Every red dot you see is a Tweet or Instagram post from an athlete since the Winter Games kicked off.

Woah, I can see Russia from my house! What we see are two main clusters, one in the northeast and one in the southwest. And those social hot spots match up with the two main geographic areas where the Winter Games are being held this year - the Olympic Park (southwest) and Mountain Cluster (northeast.) Seems like we’ll have some interesting data to analyze, so let’s dive in to each group of points. First we’ll zoom in and check out the Olympic Park.

The Olympic Park

This is the area is where the opening/closing ceremonies are being held, as well as all stadium events (hockey, speed skating, curling, etc.) And of course, there’s the infamous Olympic Village where the athletes live and hang out. What can a social location analysis of the area tell us that we didn’t know before?

Olympic Park Map

Overall, these patterns make sense. You can see athletes posting from all of the major venues, and a ton from the Olympic Village. We can also see a lot of social media activity by the Olympic Torch (located at the bottom of the large circle in the middle of the stadium cluster), and a few random posts throughout the stadium grouping just known as the Coastal Cluster.

Once we see a collection of posts in the same area, the best way to find out what’s going on in each each is to look at some individual data points for more context.

Social Kickoff: We’ll start with the big stadium during the Opening Ceremonies (remember, we also have the dimension of time to use when examining location patterns.) Here’s a post from Alex Bilodeau minutes before Canada walked out in the opening ceremony, along with the geolocation of the Tweet:

Well that doesn’t make any sense - how could he be standing in the middle of Fischt Stadium if he hasn’t even walked in with his teammates yet? This stupid thing is broken. Until you realize that the teams were all waiting to be tunnels, underneath the floor of the stadium. Ah.

Flying the Social Skies: Let’s pan up north and check a cluster at the airport. We can see a big group of athletes posting from just outside the main terminal, probably just having landed and waiting for their team buses to arrive or depart.

Airport Map

But if you look below that main cluster (red dot in the bottom-middle of the picture), we can also see Lowell Bailey (American Biathlete) posting to Twitter while still sitting on his plane having just arrived at the gate. And because these are social media posts, we can dive into the data and find out and see why he was so excited to post:

This is getting fun - LET’S DO MORE.

Ring in the Games: Here’s a great shot of some giant Olympic rings by Liam Firus, a Canadian figure skater. I don’t know what the building is behind the rings until I check the location of the Tweet, which tells me it was sent somewhere from inside one of the courtyards in the Olympic Village.

Share the Sunshine: Let’s turn our attention to a Tweet from Saija Tarkki, a Finnish hockey player, tweeting in her native tongue. After checking the geolocation of this Tweet, I didn’t even have to hit the “translate” button to know that she was talking about how warm it is outside and that the sun is shining. Why? Because even from 6,000+ miles away I can see she sent the Tweet poolside (although I’m guessing it was still a bit too cold for a swim.)


Light the Social Torch: Here's Kate Hansen, with luger Jayson Terdiman having a nice moment at the Olympic Torch. This one looks about 50 feet off - it might be that the coordinates are a bit inaccurate, or she may have posted it to Instagram after walking over to the torch.

Let’s move northeast and see what’s happening in the land of snow and luge.

The Mountain Cluster

The Mountain Cluster provides refreshing social scenery for skiing, snowboarding, and all track sports (bobsled, luge, etc.) What patterns can we see with location-based social media data mining?

Mountain Cluster Map

In a very similar fashion to the Olympic Park, posts happen at the games but more often in the supporting Olympic Villages where the athletes live. I found many examples of posts that were taken on the slopes but posted back at the Olympic Village. Remember: geolocation is tagged from where you post a Tweet/Photo, but not necessarily where you took the photo. I’d guess this is due to costly international data rates (and wifi back at the Village), and just being too busy to post in real-time all of the time. But we still have plenty of posts that happened right from the location in the Mountain Cluster. Here are just a few examples:

Selfies For The Gold: Here’s a self-shot by Aja Evans at the base of the Mogul Course. Because she had location turned on for Instagram, we can see exactly where it was taken:

Connecting the Dots: Below is a map of geolocated posts throughout the Olympic Village, but with two lonely posts at the bottom right of the photo that seem to be off with no others around them. 

Even stranger, they seem to be posted from the slopes where the athletes are competing. Certainly no self-respecting athlete is going to post during their run for Olympic glory. What are you doing on the slopes, little dots?

Diving into the data, we see that it’s two athletes posting from the same spot (a giant inflatable Russian doll) on the Slopestyle course a few days apart.

On the Social Track: And remember when I mentioned that social geolocation is accurate within a few feet? Check this out - here’s a series of Tweets/Instagram posts from athletes during the past week, mapped out by longitude and latitude:

It’s the bobsled track.

To explain this in analytics terminology, that’s cray cray. The level of detail with social geolocation gives us incredible mapping capabilities to the point where I can almost draw the bobsled track from 6,000+ miles away without the map. And to be sure, let’s check a few of the posts:

Yep, seems legit.

I could honestly keep going all day with these. Social media has become so pervasive in our society that even though a small fraction of athletes have location services turned on, there are still enough posts to find patterns with latitude and longitude that would have been impossible just years ago. And the important takeaway here is that these patterns don’t just exist in Sochi - this type of data is being created every day, all around the world.

What Have We Learned?

Location is just one more dimension of social data that gives you more insight into the conversation happening about topics, events, and brands. While you may not want to examine each individual post to understand the story of your customers, looking at location data in aggregate can help tell a story that would otherwise be hidden. Will you use it to segment social conversation and sentiment by region? Find new, untapped locations where you don’t currently serve customers? Now that you know that social media brings highly accurate location information, the answer is up to you. Now if you’ll excuse me, I’m going to call in sick and hit the pool (with my location services turned off, of course.)

Chris Kerns's picture

Chris Kerns

Chris Kerns has spent more than a decade defining digital strategy and is at the forefront of finding insights from digital data. He currently leads Analytics and Research at Spredfast. His research has appeared in The New York Times, Forbes, USA Today and AdWeek, among other publications.