Data layers

Courts

The first and most important layer are the actual courts. These are outdoor courts (i.e. the live outside a building) that live in the Overture Maps dataset which contains data from OpenStreetMap, and it appears that most of the courts in the data are actually from OpenStreetMap. You can see an example of one here.

There was a little bit of extra filtering to exclude some results like a generic sports “pitch” to really focus on the ones we know are basketball courts.

Indoor courts

But of course basketball is also played inside so we have to map those, but we have no way of knowing if there is a court inside a building so we have to guess. First which buildings are likely to have a court inside them. I made a few assumptions:

• Schools: nearly all large schools have a gym and likely have a hoop but not all schools are large enough to support one

• Churches: Many churches also have a gym or some sort of recreational space, again using size as a filter here

• Gyms: This one was a bit harder to measure but gyms in large buildings were one final layer.

I also pulled out the YMCA/YWCA listings as a separate layer because they have a unique history in the sport of basketball. I didn’t end up using it in the final weighting but it was worth a look just to see.

Now, how did I know that a court wasn’t inside a building? I looked. For example for this YMCA in Martha’s Vineyard you can see the building and a court right next to it on the Overture Maps map but it’s missing on the satellite and you can even see a pin for a basketball court on Google Maps.

Thank goodness for Street View because we can see a basketball court hiding behind the trees.

Now to be clear I only looked at 10 of these to make sure the data lined up but all 10 did so I went ahead with the assumption that we should be okay to move ahead.

A note about “gyms”

The gym category in the data is way too wide to capture anything meaningful so I am in the process of scaling that back to focus on known facilities with a gym like Lifetime or LA Fitness. Having a bunch of “fitness” gyms for lack of a better term isn’t helpful for our purposes.

Teams

Teams most certainly play a role in the sport, and a winning team can drive interest in the sport. And a team exists at a specific place, but the geography of sports is much more of a cultural or human geography.

To capture this we want to map all the teams from the NBA, WNBA, and D1 Men’s and Women’s leagues. Now I wanted to also include some metric of success here as well so I pulled in their winning %, winning % in the past 10 year, number of championships, and years in existence to help balance the influence out. The teams you would expect stand out in the map.

For pro teams I used the NBA’s 75 mile market designation and 40 miles for a college program which you can see below. For each map more blue = more successful based on the index, red = less successful.

Players

This was one I wanted to include for two main reasons. First and foremost it is a good proxy of where there are programs that support basketball development, not perfect but again a proxy and acts sort of as a metric of where people are born. But I did want it in there to help map the "Kinston, NC” effect if possible, which is home to 22,000 people and all these NBA players:

• Brandon Ingram

• Jerry Stackhouse

• Cedric Maxwell

• Reggie Bullock

• Charles Shackleford

• Mitchell Wiggins

I also didn’t tie players to teams in this case. Team location and success is already mapped so adding players in basically amplifies an existing signal. But there is one problem summed up by one of the greatest players of all time.

Michael Jordan was born in Brooklyn. So he actually counts as a data point in Brooklyn rather than North Carolina or Chicago where he had arguably the most influence geographically. The problem is there is no way to adjust or edit this in the data across every player. The player data I pulled between Wikipedia and other sources has hometown and trying to manually adjust this would take, well a really long time. So I am going to make the hometown location of players the least important factor in my analysis. Which leads me to my next section.

Analyzing the data

So now I am going to move on to the analysis of the data. Here is my plan as it stands now:

• I am going to weight the different features as such:

  • Courts: 0.4

  • Schools, churches, gyms: 0.3

    • High schools will be weighted higher than middle schools, middle schools higher than elementary due to the fact that high school teams have a certain weight to them)

  • Pro and college teams: 0.2

    • They will each have their own individual team weight as well but this helps show that they exist but at a higher cultural and human geographic level

  • Player hometowns: 0.1

    • For the reasons explained above

Now to measure this accurately I am going to create a grid across the United States using hexagons which is known as H3 cells to create even grids across the country. This will take into account all the features above in each cell using the weights as I described above and then normalizing that with population to balance it out.

And here is my favorite part of this article: introducing the two laws of geography. They were created by Waldo Tobler who was an infulential geographer sometimes refered to as a “mathematical cartographer”. These rules generally hold true across any geographic phenomena.

His first rule states:

And his second:

In short the rules tell us (in my interpretation in this specific instance) that for each hexagonal cell on the map, things nearby will be more related to it than things outside of it, and those things outside of it will impact what happens inside it.

So we will use the H3 cells to create some rings around it and the H3 system sets up perfectly to do this. From here we can use distance decay to weight the nearby cells appropriately so we can measure outside influence (respecting the 2nd law) but measure proximity (respecting the 1st law).

Now there are a lot of decisions built into this. Some of these are my decision and some of them are yours (in the first post you pointed out that courts should be the main factor). Now any change here to the analysis could change the outcome.

In this field there are no “absolutes”. Multiple places could be building a court right now that isn’t on the map. The data doesn’t capture everything between players, courts, etc. The weights are my subjective judgement. But that is what this post is here for, to lay out the decisions so you can see them.

What caught me by surprise

But the coolest thing so far about this project is the people that not only found it, but also reached out.

The person who runs the Heartland Homecourts account on Instagram sent me a message about their project to photograph and document courts across Indiana and, well the results speak for themselves and capture something this analysis can’t: the individual home courts in driveways and yards.

The team from Local Hoops also sent me their map which maps user submitted photos of courts across the country and the world (they also have incredible clothing, hats, and more so make sure to check that out too).

All of this helped to validate this project to me even more. There is clearly something tangible about basketball and geography. It ties local identity and community to something much larger, and that story is one worth looking at. More to come as I start to get through the analysis.

Let’s start with the big question: why is it even worth studying basketball in the United States?

For me it boils down to three things (pulling this directly from my Notes app as I was thinking about it last night):

1. The access

2. The lore

3. The geography

Let’s start with the access. If I was looking at any sport outside the US the sport to study would be soccer (football) because it is the most accessible sport outside the US. But in the US, Basketball might be that sport.

In fact during the most recent Finals, Kenny Smith said something to this extent, about basketball being the sport of New York City. I think that is actually pretty universal across the country. You just need a court and a ball, and one other person if you want to play a game. On a nice day you are likely to find a game or at least someone playing alone on many courts. Can you say the same about baseball or football?

It was also my suspicion that it is the most accessible sport from courts outside to courts in schools, gyms, churches, and other institutions. So I hope to answer that question.

Next there is the lore. Basketball has this gravitational pull in American sports. Culturally it has a massive impact in so many areas. Nearly all of the most followed American athletes are basketball players. But this stat stood out to me:

I had an inkling but the data tells the story there. Then there are the movies from Hoosiers, Coach Carter, Space Jam, Hustle, and more. Then there are the rise of the basketball podcasts and creators from Lethal Shooter and many more.

But the lore goes deeper to the specific courts and stadiums, from MSG to Rucker Park, Cameron to Allen, and games played in football and baseball stadiums and aircraft carriers. Not to mention we have a month dedicated to the sport.

The final one is the geography.

We have some clear geographic centers of basketball in the United States and it’s worth answering where those places actually are. There are the cities: New York City, Chicago, Los Angeles all have clear and defined basketball cultures.

But there are bigger regions too. Indiana claims to have that identity for many reasons but so does North Carolina and the density of successful college programs there helps make the case.

But we don’t have an answer. Go search for the “mecca of basketball” and you will get a ton of different answers.

So, can we answer it with data? We will find out.

What are the questions I want to answer

So that boils down to three questions I want to answer about basketball:

1. What is the geography of basketball in the United States?

2. What does that tell us about the country and the sport?

3. Is there a center (or mecca) of basketball in the US?

Now we need data to answer those questions, but drop any questions you might also want to answer in the comments.

What data can help answer that question

So far I have picked these data layers to drive this investigation:

1. Courts: all the courts that we can see in parks and public areas from OpenStreetMap and Overture Maps

   1. But also the ones we can’t see in schools, gyms, and churches. We can’t be certain but we can estimate that if a building is over a certain size that a gym is inside and a confidence level (i.e. most schools will have a gym with a hoop if they are over a specific size, and the more senior the school the more likely)

   2. This will also need to be balanced by population density as well as access in different socioeconomic measures too.

2. Gyms: given the prevalence and importance of gyms specifically the YMCA I want to give these an extra weight in the model

3. Teams: having a team from the professional to the collegiate level will drive some level of importance in those communities. We can also weight those teams based on their success and winning percentages as well.

4. Players: players that have made it to the collegiate and professional levels also gives us a way to measure how much people are participating in the sport in different areas for both women and men.

Those are the four main areas I am focusing on. Of course this is the point of this article: are there any that are missing, and which are most important?

Drop your thoughts in the comments and let me know in the poll which layers are the most important:

Welcome to Ground Truth. The goal here is simple: to analyze trends that aren’t immediately apparent, and get you to help me as we go.

Every map reveals some specific layer about how we live. One of the reasons I love doing this work is because even though we have mapped many things, there is a lot left to uncover.

Maps and geography are exciting to a lot of people because they help uncover ways we live and interact with the world, but they also inspire a sense of exploration. It may seem like there isn’t a lot left to uncover, but I don’t think that is the case, and I hope to show that here (and get your help along the way)

I’m Matt. I’ve spent 16 years working with geospatial data, and I make videos about maps and geography.

Here I will be running what I call “investigations”. I did my first one alone (the most isolated building in the world) but the comments gave me way more ideas so I wanted to find a way to open that up to more people.

So, an investigation is a real question, where we will try to answer it with real data, and every step published as it unfolds. You see the messy datasets, the dead ends, and the judgment calls I have to make along the way. You also get a say in those calls.

Not every investigation will have a concrete answer, but that is part of the process and I hope you will participate in that process.

The first investigation I want to look into: the geography of basketball in America.

Why basketball? It’s one of the most accessible sports we have in the United States. All you need is a ball and a court. There is a gym in most schools and many other buildings (gyms, churches, etc.) all over the country. Basketball is part of our cultural fabric in film, TV, clothing, and more. Not to mention every March one of the largest sporting events takes over for a month.

So I already started getting the first layers. I pulled all the tagged courts in the country from Overture Maps data, then layering in schools, churches, and stadiums (and more) until we have the full picture of where the game actually lives. By the end I hope we will have the number, the footprint, and the places: the most hoops-centric (centroid) place in the country, and the farthest you can get from a court in the lower 48.

The first decision: what counts as a basketball court? A church gym. A school playground. A driveway hoop (which I would love to map but nearly impossible and we will look into why). You’ll help me decide, and the ideas you set will drive what we look at.

Founding readers help to drive the analysis. My plan is to publish twice a week (as often as I can). First one, Field Report, for free subscribers. These are the notes and results of what we are finding as it happens. Second the Lab Notes for paid subscribers. I’ll show you the state of what I found, where we can go next, along with a poll on key decisions to drive what comes next.

So subscribe to follow along.