How Vox visualized segregation in the United States

Behind the scenes, Interviews
Share on FacebookShare on Google+Tweet about this on TwitterPin on PinterestShare on LinkedInEmail this to someone

Vox data journalist Alvin Chang, an alum of the Boston Globe and ESPN, has been writing stories about segregation for most of his career. Earlier this year, Chang looked at segregation in both the workplace and the home in his story “American segregation, mapped at day and night.”

Using data provided by researchers at Cornell and Penn State, Chang created an immersive story with interactive and visual features that blended historical footage and three-dimensional visualizations to illuminate just how segregated America still is.

Storybench sat down with Chang to talk about how this story came about and the steps he took to create an immersive piece.

How did you come up with the idea for this story?

I’ve written a few pieces on segregation. It’s often about racial segregation in housing and in schools as those are kind of the two biggest examples of American segregation that was engineered by the federal government. So, I’ve done quite a bit of work – especially in the school segregation space and especially in the mapping and the analysis of this data.

[For this story,] I actually got a DM on Twitter from one of the researchers – his name is Matthew Hall – and I think I cite him in the piece. He was like ‘Hey! Me and my colleagues have this really interesting piece of research. I’m curious if you want to look at it?’ and I jumped on a call with him and I said ‘Yeah, definitely! Especially if you’re willing to share some data with me.’ So, he actually reached out with some research he had and a lot of the time these researchers have a lot of data behind their papers but their data is stuck behind, essentially, a PDF or an academic journal and no one really gets to see what their findings are and so when I see examples of this I jump on it. I’m like ‘OK. I definitely think these people did work that is incredibly important and my job is to try to help readers understand the findings and maybe personalize it to their actual geography or their situation.’ As soon as I realize this was one of those projects – I talked with Matt, made sure that everything that needed to be agreed upon was agreed upon, and then he shared his data with me. From there, I just started mapping and a lot of the times the narrative that journalists want to tell isn’t clear. It’s not clear from the research paper, you kind of have to dig through to figure out how to tell the story more clearly and so one of the things that I realized I needed to do is call this ‘segregation in day and night;’ versus what they look at work versus at home. You know, most people are at home during the night and at work during the day so the idea was if we could give this a time element it might help people better understand what they’re looking at.

Once I started mapping this and gave myself this idea of what I’m looking at, I started finding some really amazing patterns. And these are things we kind of already knew but, especially when you look at black Americans in any major city, when they’re at work they’re in a much more diverse environment but then they go home to highly, highly, segregated neighborhoods. Same for Hispanic-Americans in many cities. Asian-Americans less so but especially in L.A. there are some really interesting data and a lot of Asian-Americans go home to highly segregated neighborhoods. And so as I said ‘OK, there’s something here. There’s something that really resonates. Something you wouldn’t see otherwise. Something that you kind of know in theory about the way that segregation works but I had never been able to see it this way on a map and in the cities that I lived in.’ For example, I grew up in Kansas, so the first thing I did was I went to Kansas City on my map and I started seeing where black people work and where do black people go home and it lights up red in certain neighborhoods and I thought ‘OK. I knew this in theory. I knew that at a workplace – which is relatively diverse – it feels like all of us come from similar backgrounds and have similar lives. But it turns out the color of our skin still matters a great deal,’ and you could see that based on where we’re about to go home after five o’clock. That’s how this kind of came to be and from there it was a matter of making the visualizations make sense, writing the story, and then starting to write a video script that I could do a video version of – which is kind of a different story than the actual text version but it uses the same data set.

How did you pitch it to the newsroom?

I pitched it to my editor with basically the exact same headline. [I said] ‘I think I can do a map of segregation during the day versus night’ and, you know, I’ve done enough of these pieces on residential and school segregation that they said ‘OK. We trust you to turn it into something.’ That was kind of the start of that. But when I first pitched segregation stories, I needed to have a really precise angle – a really exact angle – that will add some value. If I was just like ‘I’m going to show you segregation is worse at night versus the day,’ it’d be boring. It wouldn’t mean anything. But this is going to show you exactly how it is in your town, your neighborhood, and I think that’s kind of what got the editors interested.

Were there other aspects that you and the other researchers were considering? Did they ever consider public space like malls compared to homes or malls compared to sports centers or other places that everyone goes to?

If you think about this in terms of where is everyone going the day versus night – everyone lives through the day versus the night, right? However, it’s not if it’s a question of the mall but it’s ‘What does that exactly mean?’ I think the question is ‘How segregated are certain spaces?’ and I think it’s a slightly different question than when we’re looking at different spaces.

I think it could work but what these researchers were interested in was segregation at work and how that changes versus the neighborhood that the workplace is located in. I think another aspect they could look at is segregation in public spaces and other people have done work throughout public spaces but it becomes much harder because it’s like ‘Where exactly are we going to find that data of people?’ Actually, I believe – for the Boston area – there are some researchers that have done this work. But they’re basically trying to look at public spaces and how segregation changes in those public spaces so I think the question is a little bit different.

How long did it take to actually compile it and make the map in the story?

The map itself took about a day but it’s making the last 10 percent [of the map] work like getting the transitions to work really well because you can notice when you change from work to home there are big areas that animate from one to another. That took a little time to figure out. I kind of had to hack through this mapping software to make that work. Another thing is when you first get to the map it gives you a little guide where it starts and it gives you like ‘This is where white workers are in your area at work’ and it was colored by how segregated it is. That kind of narrative took a little time to develop but the initial mapping portion took about a day to get everything on the map.

Chang’s map allows users to select from dozens of U.S. metropolitan areas.

Did the writing of the story come before or after the maps?

After. Once I knew what was on the map, I can go back and write the story and then a lot of the story is just giving context to the personalized data that you’re going to see on the map. That’s kind of what the text story is about. The video – a lot of it – is about giving historical context about what you’re about to see.

Why did you choose not to put the historical context into the story?

I think part of it is that I had more time to work on the video and on top of it – a lot of the time text pieces are not necessarily read for experience in kind of a narrative format, especially when there’s an interactive there. Right now, you get to this map and chances are there are lots of people that aren’t even reading the first few paragraphs – they’re just going to the map. So going to the map and playing around with the map they’re like, ‘Oh! This is really cool’ and then they might be leaving the page. Or, what they want is another way to look at this dataset.

I believe I put a little bit of background in this piece but it ends up being a thing that is based around the personalized interactive and if you want context – you could get it elsewhere. But the video alone could give you everything you need to know about what’s happened. I’ve covered this enough times that I can link them [to another story]. If I end up putting the context in every story then I’m just copy and pasting the same paragraphs over and over again – which is fine, there’s nothing wrong with that but I think the editors in this case preferred this piece be tighter and be focused on just around the interactive. The video, meanwhile, focuses a lot of the times on how people experience videos. They’re on Youtube, playing around, and they think, ‘Oh! This video looks interesting’ and clicks on it. If all they saw was how segregation changes but there’s no context to it – if there’s no ‘here’s why this matters’ – then I think you kind of feel lost. And Youtube, it’s a separate platform. You might have been watching a video game streamer before that so you need to add some context about what you’re about to show.

Can you walk me through a step-by-step from when you got the data to when you finished the map?

The first step is figuring out how the researcher is going to get you the data. I believe he gave me a Dropbox so I downloaded all the data. Then, from there, it’s a matter of figuring out what unit of geography the data was in. Right now it’s divided by census tracks I believe. You kind of have to figure out how the geography is divided up. Once you figure out it’s by census tracks you need to download the shapes for every census tracks in America because otherwise how would the program know which shape the geography is. Once I get the data,I know the geography, and I download the shape files for the geography – I link up the data to the shapes. That’s the first thing.

From there, I organize all the files so that when I tell the computer program ‘Hey, go look for everything in the New York metropolitan area’ I know which file to look in. I know how to tell the computer program where to look for that file – the file being the shape file with all the data. Once I organize all the files, let’s make a map with just one of the metropolitan areas – let’s say it’s just New York. We do it with just one of the metropolitan areas and you get it to work, then all you have to do is say ‘Now that one of these works, lets direct you to another file. Let’s go to Chicago now and see if that works’ and then Chicago works. From there, I made that little drop down menu and then that lets you automatically – instead of me telling the computer program where to shift the map to – can have the users do it themselves so I can shift to New York from Chicago. At that point it’s just pretty self explanatory. You just need to map whatever colors you want to put on the map.

Was there a certain tool you used?

The base map – the fact that the United States is behind here and it looks like a real map – that is from Mapbox. Then, the actual program that helps turn the code into a map, is called Mapbox TS and that’s like a kind of library that helps connect and helps me write code so I can map all these little shapes into 3D things.

Do you have another overarching message that you want the readers to understand from this piece besides that segregation still exists prominently in modern day?

I mean, I think that’s the overall takeaway. But I also want people to go from one place to another, virtually every city, and see that it exists everywhere – not just in their city. That’s just the scale of it that really gets to you like ‘Woah! This is everywhere.’ So, that’s it.

For someone that wants to make a map themselves like this and a story to match – what tips can you give a journalism student interested in doing a piece like this?

I think the most important thing is to know that there is a story in the data or have a sense that you can find a story in the data and what that story is doing to be. It’s essentially news judgement but it’s not news judgement in the same way that newsrooms are going to talk about it like ‘Is this a story or not?’ It’s more so ‘How much do you know about these topics? How much context do you have?’ so when you see hints of a story you can extrapolate that out and say ‘Yes, that is a story because of X, Y, and Z’ as opposed to just ‘Oh! That’s interesting. People will be interested in that.’

I think it doesn’t quite work when you’re working with big datasets that are coming from researchers because usually when you look at it you think ‘No one is going to look at that – it’s boring.’ But then you kind of have to dig a little further using the contextual knowledge you have to say ‘I know why this is important’ based on the contextual knowledge that I already have. I think having done a lot of research and reporting is crucial to be able to build something like this first and foremost. And I think from there it’s a matter of designing out the experience you think people want to have and figuring out the technology. That ends up being slightly easier because you know what you want to make but first you have to know what you want to make.

When you make pieces like these do the researchers usually reach out to you or do you find your own data?

It’s 50/50. Usually, I’m finding my own data but sometimes researchers will say ‘Hey! Can we have a conversation? I think I have something interesting here.’ We’ll kind of collaborate to make sure I can have their data and they can explain the context behind this.

When did you start becoming interested in these sorts of subjects and stories?

I think I’ve always been pretty interested. I’ve been a data reporter for a while now and when I was in Boston at the Boston Globe we did quite a lot of work on segregation in the modern day and segregation, violence, and concentrated poverty. But I think as I started learning more and more about the history of why this exists, I realize that it’s important to show that segregation exist in our everyday lives even in modern day and for people to be able to see that clearly. Because I think that it’s so easy to go about our lives and think ‘Oh! It’s not that bad’ or like ‘Oh! This is because like people of different races don’t want to live with each other so that’s why segregation exists.’ It’s so easy to think like that and I think tools like this and experiences like this and stories like this can hopefully help people realize that ‘No. It’s actually an engineered demographic pattern.’ This exists because someone wanted it to exist like this.

Audrey studies journalism at Northeastern University.

Leave a Reply