Jeremy is the author of American Inequality, a biweekly newsletter that uses data visualization to highlight U.S. inequality topics and to drive change in communities. His work has been published in TIME, Bloomberg, and the LA Times. He was a dual-degree masters student at MIT Sloan and the Harvard Kennedy School and was formerly a macro policy strategist at the Federal Reserve. He now works at Google and lives in Brooklyn.
Episode Notes
Jeremy on Twitter | Op-ed in Time
American Inequality newsletter: americaninequality.substack.com
Federal Reserve Bank of New York
Technology and Disability: The Relationship Between Broadband Access and Disability Insurance Awards
Some coverage of the map:
Related Episodes
Episode #228: Ethan Mollick
Episode #224: Pieta Blakely and Eli Holder
Episode #191: Sarah Williams
Sponsor
Use my special link (https://zen.ai/policyviz12) to save 12% at blendjet.com. The discount will be applied at checkout!
New Ways to Support the Show!
With more than 200 guests and eight seasons of episodes, the PolicyViz Podcast is one of the longest-running data visualization podcasts around. You can support the show by downloading and listening, following the work of my guests, and sharing the show with your networks. I’m grateful to everyone who listens and supports the show, and now I’m offering new exciting ways for you to support the show financially. You can check out the special paid version of my newsletter, receive text messages with special data visualization tips, or go to the simplified Patreon platform. Whichever you choose, you’ll be sure to get great content to your inbox or phone every week!
Transcript
This episode of the PolicyViz podcast is brought to you by BlendJet. I’ve had my BlendJet blender for about a month now, and I love it. Instead of grabbing a bag of chips or a cookie or a brownie or whatever’s leftover on the counter, I really quickly, really easily blend up a little smoothie in the afternoon, gives me that little afternoon perk that I need, filling me up to get me through the end of the day. The BlendJet 2 is a portable blender, so you can blend up a smoothie at work, a protein shake at the gym, or even a margarita on the beach. It is small enough to fit in a cup holder, but it’s powerful enough to blast through tough ingredients like ice and frozen fruit with ease and it is whisper quiet, so you can make your morning smoothie without waking up the whole house. The BlendJet 2 lasts for more than 15 blends and recharges quickly via USB-C. And best of all, BlendJet 2 cleans itself – you just put in some water, a drop of soap, and you’re good to go. So what are you waiting for? Head over to blendjet.com and grab yours today. And be sure to use my promo code, that’s POLICYVIZ12 policyviz-one-two to get 12% off your order and free two-day shipping. No other portable blender on the market comes close to the quality, power, and innovation of the BlendJet 2. They guarantee you’ll love it or your money back. So blend anytime, anywhere with the BlendJet 2 portable blender, go to blendjet.com and use the code POLICYVIZ12 to get 12% off your order and free two-day shipping. Shop today and get the best deal ever.
Welcome back to the PolicyViz podcast. I am your host, Jon Schwabish. I hope you’re well. Spring is well underway here in Virginia, the pollen is everywhere, we tend to call it the great pollening here, so eyes are itchy, noses are running, but we soldier on. I’m excited for this week’s episode of the show. Before we get into that, I want to let you know that my next book Data Visualization in Excel is about to hit bookshelves. This is a step by step book to help you create better and more effective graphs in the Microsoft Excel tool. I’m a big believer that any data visualization tool can be used for different purposes. It’s not one tool to rule them all, this ain’t the sort of Lord of the Rings data visualization tool out there. So I use Excel for a lot of my work, I don’t use it for all of my work, because it’s not going to work for certain types of visualizations, or certain products, or certain use cases. But if you want to expand how you use Excel, you want to create better, different types of graphs, I hope this book will be the book for you. So I’ve put a link to both the CRC Press page, which is the publisher of this book, and to the Amazon US site, and if you’d like to go and preorder it, of course, as you probably know, preorders really help, get the word out to more and more people so that they can use the book in their own work. So I hope you will check it out, that’s Data Visualization in Excel hitting bookshelves any day now.
Having said that, on this week’s episode of the show, I chat with Jeremy Ney from americaninequality.io to talk about a recent map he created on disparities and life expectancy across the United States. And it was a map that got picked up in lots of different places, including the New York Times, including the Washington Post, and so, Jeremy came on the show to talk about why he thought the map did as well as it did, his ongoing work and the different tools that he uses in his own process to analyze, collect and visualize his data. So I hope you’ll enjoy this week’s conversation with Jeremy Ney, here on the PolicyViz podcast.
Jon Schwabish: Hey Jeremy. Good afternoon. How are you doing?
Jeremy Ney: Hey Jon, I am doing very well. How are you?
JS: I’m doing great. Thanks for getting in touch. I’m excited to talk about this work you’ve been doing, and all the places you’ve had to follow up. It sounds like you’ve been busy with this particular map that we’ll focus on.
JN: Yes, very busy, and a lot of interest in life expectancy and the internet, yes.
JS: Yeah, it’s really interesting, and we’ll get to it, but really interesting how you make that like one thing, that kind of just takes off all of a sudden. So I thought we would start maybe you could just tell folks a little bit about yourself, and what you’ve been doing, and then, we can segue into the current work that you’re up to.
JN: Yeah, absolutely. So I am the author of American Inequality, which is a newsletter and data portal that uses data visualization to highlight US inequality topics that are often left in the dark, things like life expectancy, internet access, food deserts, debt, things like that. This work really emerged from a lot of research that I was doing when I was at the Federal Reserve in New York, mainly looking at US income inequality, and through that recognized that inequality was about way more than just income; it’s tied up in education, and healthcare, and taxes, and race and gender and location. Went off to do more research at MIT and Harvard, where I kind of started building out more of these data visualizations, was sitting on them and decided to turn it into this publication, which is sort of what we’ve now been digging into. And then, I also worked at Google, doing tech policy work there as well.
JS: Cool. So let’s talk about the map to start, the most recent map. I don’t want to call it the map, because you’ve done a lot of different things. But I think our conversation will branch back into some things you just mentioned, because I’m curious about your experience, both at the Fed, and then, at school, and how DataViz sort of played a part in all the different roles. But maybe we could start on your life expectancy map, so this is a choropleth map, county level of the US, looking at disparities in life expectancy, and that’s all I’m going to say about it. And I’m going to let you explain it for folks who maybe haven’t seen it. And, of course, I’ll put it on the notes page, so they can take a look in more detail, but I’ll let you take a spin.
JN: Sure thing. Yeah, so the biggest thing that we kind of found in this map, and I think why it caught so many folks’ attention is what the data shows is that the US is actually experiencing the greatest divide in life expectancy across regions in the last 40 years. So what you can see from some of these red spots, or blue spots is that if you are born in certain counties in Mississippi, or Florida, you may die at 67 on average; but if you’re born in certain parts of Colorado or California, you may live to 87 on average. So a 20-year gap in life expectancy from no fault of your own, you may live there or be subject to a lot of these challenges of that region. And so, I think that that was what was really quite striking to folks in this is that when we talk about averages, or the average experience of America, we’re actually not digging into those communities, we actually see much bigger divides, and not everyone is really experiencing those same outcomes, that can be driven by a whole host of factors within those communities.
JS: I’m curious because there’s been a lot of discussion about changing life expectancy, obviously, across nationally, but also within the United States, you’ve seen real dramatic shifts in mortality, particularly among middle aged white males, which sort of reversed the decades’ long trend, and I’m curious why you think this particular map struck such a chord.
JN: Yeah, so I think there were three things that were kind of happening all at once to create this perfect storm of why this really took off. So I think, first and foremost was John Burn-Murdoch at the Financial Times published this piece talking about US versus UK life expectancy, basically showing that Americans die far younger than British people do. I think a lot of the data visualizations that John presented were quite striking to folks; and, in particular, that the average American actually has the same life expectancy as the worst region in the UK, in this area, Blackpool. And so, I think that really drew this interest in life expectancy kind of into the fold. I think as people started to dig into it, they’re like, hey, the average of the US might not be quite right, because there’s such different experiences that we can have in Alabama versus Washington State, for example. So that was number one.
I think number two was new data just came out basically around the same time. And so, the new data there was that the US actually hit the largest decline in life expectancy in the last 100 years, much of that was COVID driven, but some of it was these other factors as you had mentioned. But basically, not since 1921 to 1923 had we seen the same terrible decline. And on top of that new data showed also that the US is far, far, far worse than any other country, when it comes to some of these declines. And then, third and finally on kind of this perfect storm is that there’s so much that we talk about in the US about these different divides that we have, whether it’s political divides, or healthcare divides, or race based divides, and so, I think there was something about life expectancy that really allowed everybody to come together, and sort of have this appreciation for this common challenge. We all want to spend many years with our loved ones. We want to celebrate birthdays with each other.
And even as the map was blowing up, this one woman reached out to me, this woman, Amy in Kentucky, who talked about how her father died at 46 from cancer, and how the only doctor that he ever saw was the doctor that he saw in prison. Now, she was so worried that she herself wouldn’t live to 40. And so, I think it’s these types of challenges that everyone is so worried about like their longevity. And Kentucky too has the fifth – I think it’s the fifth worst life expectancy of any state in the US too. And so, I think it’s this new work that is being done comparing the US across countries, it was this new data, and it was this new, like, this continued sense of feeling that a life is really precious and I’m so concerned about it that really allowed I think this DataViz, in particular, of the many that we have to really take off.
JS: Yeah. I think, personally, I would add a fourth to that, and that the structure of the map itself was really interesting. Because if you said to someone create – someone in the DataViz community said, create a map for me, here’s the data, create a map that shows life expectancy, the range is from about 66 to 90, I think most people, when they create that map, they’d create it using a sequential color palette. It would go from light blue to dark blue. But the way you created it, was you did it as a diverging palette, where the center point is the national average. And so, you’ve got this above average in shades of blue, and below average in shades of red, and so, I think in a lot of ways that draws a pretty stark difference that is really visual and sort of instinctual on the map rather than just going from like a light color to a dark color.
JN: Yeah, I think the two differences in colors, the blue versus red is really striking, not only for drawing your attention to those areas in red, and kind of that signaling of when we do see red, we tend to think of some sort of calamity going on in some ways. But what we have been working on quite a bit more at American Inequality is this concept of opportunity mapping is how can these areas in red that are particularly struggling with a challenge, like, learn from those areas in blue as well. And so, I think when you have those diverging palettes, it’s easier for you to see, like, hey, I’m actually in one of these struggling regions, how can I better identify someone else who is doing better with this challenge. So it’s finding those comparable regions to say, hey, I’m maybe challenged with this particular inequality, can I find some other area that can help me work through it.
JS: Yeah, you mentioned the politics of the kind of our divided country too. Do you think like the red/blue, do you think it sort of harkens to that Democrat/Republican color palette? Do you think people try to mentally make that link?
JN: I think people definitely make that link, and, in particular, the day or so after the map really blew up, Paul Krugman wrote an op-ed in the New York Times talking about exactly that, where he talked about this actually relationship between Democratic and Republican states, and in particular counties that swung one way or another in 2016 and in 2020 as well, and their relationship as well with life expectancy. And so, I think that does I think you’re right that that color choice in some ways does work into that. And then, Paul Krugman, basically went further to say, hey, in fact, we actually do see a bit of this.
JS: Right. So you grab new data, and the map is made in Datawrapper. And so, I’m curious about your toolkit – I mean, you’ve talked about working at Google, working at the Fed of New York, MIT, Harvard – and so, I’m curious about your toolkit, both from the data extraction and cleaning process, all the way through the end, through the visualization. So for you, what does that look like, and has your toolkit sort of changed over time?
JN: The toolkit has definitely changed over time, and I will say Datawrapper is an absolutely fantastic tool. I cannot recommend it highly enough. I think particularly for folks who are starting out earlier in their DataViz journey, it’s such an easy tool to create those interactive tools to help start building narratives, to help give you prompts as well about what works. I think it’s really quite helpful there, and it’s also been nice for helping to embed your work in other tools like Substack or things like that, where we have our newsletter. But our processes really, you know, we often, we will either start with an idea, like, hey, we really want to write about child care and inequality or something, or we’ll start off, you know, someone will send us an interesting dataset.
And so, some of the other tools that we’ll use, depending on how big the dataset is, there’s a lot of Python and R as well to access those from large government agencies like the CDC or the EPA or the FBI to try and pull that in and clean that data. But we also are always really trying to focus on these county level maps, in particular, because inequality happens in communities, right? And so, having state level or national level of data is helpful, but it actually can obfuscate what’s really happening on the ground, particularly, if you look at some of these bigger states, like, California or Texas or Florida, where what’s happening in the north or the south can really be quite different.
And so, whenever we’re going to those big government agencies, we try and find that county level insights, and then, depending as well on what we’re trying to build out, Tableau has also been a really fantastic tool for creating some of those dashboards to say, hey, I am a state or local politician or policymaker, I really want to understand what is this factor in my community that’s most strongly correlated with life expectancy, for example. And so, you create all these interesting toggles that we built a tool around that as well to say it might turn out that it’s cancer is one of your biggest factors or gun violence or something like that, that Tableau is quite helpful for understanding some of that.
JS: Right. And to your point about the county level, there’s another graph in the original, I guess, the original medium posts, that maybe hasn’t got as much play, but is an interesting visualization that really, I think, takes advantage of the fact that you have a lot of data. I mean, you have county level data for, I don’t know, there’s like maybe about 10 different years since 1980. So could you talk about that, like, what was the impetus of – and I will let you describe it – what was the impetus behind that second graph, and the process of creating that one?
JN: Yeah. So I’m very glad you brought that up, so there’s a handful of other visualizations that we did in there, one is this one that basically shows how life expectancy has changed over time, because I think it’s also important to acknowledge that the US has made a lot of really great strides in improving life expectancy over the last decade, and even decades before that, even going back as far as 1941 when we have some of this data from. But when we look on the county level, we actually see some regions that really aren’t making as much strides as one would expect when we talk about the overall American experience. And so, I think that’s why it’s helpful to look at that.
And then, one of the other kind of critical pieces that we found in this work is the high correlation between life expectancy and income in America, and that this is really one of those driving forces too, I think that county level data really shows that that you happen to be born in one of these counties like Loudoun County in Virginia that has the highest median household income. You are far more likely to live much longer than being born in Oglala Lakota County and South Dakota where a huge portion of folks there live in poverty, median household income closer to 35,000 or so. And so, really, this relationship that we see between wealth and life expectancy is quite strong, and having all of these county level data points really allowed us to show that relationship.
JS: What is your approach to this generally of visualizing and talking about inequality, because there are huge literatures, right? I mean, economists and sociologists and political scientists have made careers on talking about inequality and figuring out different ways to decompose inequality and different factors and controlling for age distribution and education distribution. So how do you think about that because the way that you are describing it, and the way that you are presenting it, is really here’s the data, sort of, you’ve broken down to, I mean, I don’t know if it’s a smallest level of geography, maybe the smallest meaningful level of geography at the county level. So how do you sort of try to maybe think about that or thread the needle when it comes to these different aspects of data and modeling?
JN: Yeah, I think it’s at least two parts to it. One is, we really try and just let the data speak for itself and capturing it from these large government agencies, and presenting it almost as is. We do some data cleaning, or sometimes will show these relationships, but in my experience, both from the Fed and some of the Grab research as well, like, a lot of this fancy econometric work can sometimes feel inaccessible to a larger audience. And while it’s really important to make sure you’re not falling into some of these fallacies of base rates and things like that, you do want to make sure you’re controlling for certain things, presenting the data as is I think allows folks to really say, okay, this is, I’m looking at it from the source, and I can understand it, and this hasn’t been manipulated. And I think in a time where we’re all really searching for truth and understanding what’s out there, having just presenting the facts as is, I think really goes a long way, and it’s been part of the reason this is really caught on with a lot of folks.
And then, I think the second piece why we also try and present the data in that way is it also helps us understand the stories behind the data. Right? So once we visualize it, and we see this county in red, we’ll say like, hey, what’s going on there, why is this region really struggling with this challenge, instead of doing a lot of manipulation. So, for example, when we wrote a piece on cancer and inequality, we found this one county, Elkhart County in Indiana was dying of cancer, much higher rates than in the rest of the US, couldn’t really understand what was going on, like, pretty racially diverse area, relatively high income levels, fairly high educated.
But once we basically started doing this research, we found that Elkhart County is the RV capital of the world. It’s where basically air streams are produced in all these RVs. And so, all the folks who are working and living around there are basically breathing in all of this fiberglass all the time from these RVs that are being produced, and, as a result, folks there are dying of lung cancer at four or five times the rate of the US. And so, I think that process also allows us to not come with too many priors about the data, but just to see what it is, and then, start digging in and finding the stories of the communities where the individuals are really struggling with that data. Because that’s also a really important part, right, like, recognizing that these challenges aren’t just about this test, there’s very real people that struggle with these very real challenges.
JS: So how do you go about doing that then – do you start making phone calls, do you talk to journalists, like, what is the next step in that? Because the storytelling piece, I think, we could have a whole other discussion about DataViz and storytelling, and how those words match up, but you’ve mentioned stories and community a bunch in this discussion, so what is that next step for you so you identify this county in Indiana, did you start making phone calls, did you start talking to people, like, what is that next step for you?
JN: Yeah, it’s a handful of things, I mean, part of it will be from other work that I’ve done in the past either through startups that I have worked on or areas where I’m like, no, teachers in the area, or folks who are working on startups in that particular area of inequality will reach out to them to try and get stories, either from them or from clients or something that they’ve worked with has been one area. But then we also, as you had mentioned earlier, there’s so much written on so many of these inequalities as well. And so, if we can reach out to some of the local newspapers in a region that have been really like covering flooding in an area, nobody rarely seemed to be digging into these stories except for these local newspapers, and we’ll talk to them, and try and use some of their stories about the gas station that ended up getting flooded in a 100-year flood that happened five times in the last 10 years, that type of thing.
JS: Yeah. And my guess, so correct me if I’m wrong, I’m going to guess here that throughout your education, being primarily a quantitative person, you probably haven’t had a lot of qualitative methods training formally?
JN: Yeah, I think that’s largely true. I’ve worked at IDEO actually for a bit as well. And so, that I think also helped give me kind of this deeper insight into user design and user research as well. And I think that’s also part of the reason that at American Inequality, we tend to really try and bridge those two gaps, so not always being so quantitative in the work, but trying to drive some of the qualitative as well. And we see that also, like, kind of flipped in many scenarios too where, in the early days of the pandemic, for example, policymakers knew qualitatively that they had this internet access issue in Oregon, for example, but they didn’t know quantitatively where to actually be making change. And so, for folks who could help us kind of balance that, like, we have the data, they have the stories, we’re able to actually create these great mixes. And so, in Oregon, we’re able to work with those folks and distribute hundreds of Wi-Fi hotspots with a local nonprofit across the state to really drive that impact. But, as you mentioned, it takes both of those sides for sure.
JS: And I ask because I think a lot of people who are in the DataViz field or maybe quantitative researchers, for those who don’t just like poo-poo qualitative work, who actually want to do qualitative work, it sounds interesting, it’s just – I think it’s interesting to me that your approach is to call on journalists and call on newspapers and find those stories. So if there was someone else on this call right now lurking behind me, which hopefully no one’s lurking behind me, but another quantitative person lurking behind me and wanted to do something similar, what would your one or two pieces of advice be to go find those deeper stories?
JN: Yeah, I have really found Twitter is surprisingly fantastic at that for, like, putting these stories out there, Twitter and Medium too. People really, you know, the maps resonate with them, and they’ll write me back things saying, like, hey, I’m from the only Blue County, my county doing better in my state, like, I’m so happy that I’m going to be like not struggling with this inequality. Or someone else being like, oh darn, I now realize that I’m going to be dying 10 years younger than my peers right across the border, or something like that. And it’s also been quite helpful to engage folks. The article that we wrote on internet access, someone in North Carolina reached out saying, like, hey, this article really resonated with me, because I only have one Telco in my region, they charge me exorbitant prices, and I get terrible internet service. And so, helping, you know, putting the work out there, trying to allow that stuff to come in, but having that inbound only works after a certain point; the outbound as well, I think, in particular, there’s those local newspapers, but I think Vox and the Atlantic, in particular, do tremendous storytelling work around a lot of these issues.
JS: Before I let you go, I want to get back to one more DataViz thing, and there’s a kind of a two part question here, a technical question and a sort of a general question. So the general question is: the graphs in the Medium posts, and again, I’ll link to these for folks who want to dive in more, and they obviously should – they are native Datawrapper graphs, so they are interactive. And then, on the Substack newsletter, they are static. And so, that’s kind of a two-part question. The first question is more conceptual is: why interactivity, and is that sort of like, is that a priority for you when you go in to create these denser visualizations to enable the interactivity? And then second, I don’t exactly know how to ask this question, but it’s kind of like on the technical challenge, like, do you wish you could put an interactive graph in the Substack, and do you see that as being, would that be valuable, or is it just, it’s too heavy for email for a newsletter, and it’s better to just push people back to the original where they can interact with it? So a variety of pieces in there, but it’s really focusing on the interactivity versus the static.
JN: Yeah, the interactivity is really important, I think for this work, for helping people understand the challenges in their communities. People love to say like, hey, I’m from Kings County, I really want to understand what is going on in this region in Brooklyn. Because people, the map is such an easy way that’s going to be really accessible for folks because everybody knows where they live or where they grew up, and it allows you to actually feel like this work is really touching you in a way that a bar chart or a dot plot doesn’t. And for us where we talk across such a huge range of topics as well, the map is that uniting force that connects a lot of that research and writing. So the interactivity is really important, I think for like helping folks understand what’s actually happening in their communities.
People on the Substack part, do I wish that they had the ability to do this? For sure. I think email, it’s this interesting time where I think folks are figuring out these best ways to be communicating with each other, and particularly in the PolicyViz and DataViz world and following a lot of writers and journalists now on Substack too. I see lots of folks experimenting with different ways to communicate this, like, whether it’s through video embeds, or GIFs, or static images. But in every Substack posts that we do, we put this button below every chart that says explore the data here, in part, because it helps us achieve that goal of allowing people to interact with it, but two, because we are also really big proponents of open source data in a lot of this work. Right? So not only do we allow you to access it there, but on our website, americaninequality.io, we allow you to basically download everything as already cleaned CSV Excel file, so you can access it yourself, because navigating a lot of these government portals, for example, can be quite challenging, and we don’t want that to be a barrier for folks for unpacking inequality or understanding what’s happening in their community. And so, that way of trying to lead folks to those areas, I think, also helps with that open source ethos.
JS: Yeah, absolutely. And it is worth noting that on your Datawrapper charts, folks can go get the data directly in the Datawrapper chart, which should be known for those who don’t know, like, you can turn it, like, a producer can’t turn that off in Datawrapper, like, you have the ability to turn that off, but you have fortunately left that available, so you go look at this really cool bubble chart of income and life expectancy, you want to download the data and explore it on your own, you’re able to do so which is great. Jeremy, so my last question is, where can folks find you? Where can they sign up for your materials and get more of the Inequality content?
JN: Yes. So as we chatted about the newsletter, you can sign up for it, it’s there at americaninequality.substack.com. We send about every two weeks or so. Our next one coming up will be on childcare and inequality, which we’re quite excited about. The Department of Labor, for the first time ever, just released county level data, so we’re excited to be putting work forward on that. americaninequality.io, as I mentioned, is where we host kind of all of the maps and the open source data as well, and then, folks are also interested in following along on Twitter with a lot of this work on that @jeremybney on Twitter too.
JS: Perfect. Jeremy, thanks so much for coming on the show, appreciate it, great work, love it. And I hope folks will sign up, so thanks a lot, appreciate it.
JN: Thanks for having me, Jon.
And thanks everyone for tuning into this week’s episode of the show. I hope you enjoy that. I hope you will head over to Substack and sign up for Jeremy’s newsletter. I also hope you will check out the new book page for my new book, Data Visualization in Excel on policyviz.com. And also, you can go preorder the book at Amazon, CRC Press, or your favorite bookseller. The book comes out May 26 is when we will start shipping. So I hope you enjoyed this week’s episode of the show. Be sure to check back for our next episode coming up in a couple of weeks, where we will once again help you think about ways to improve the way you create and communicate your data. So until next time, this has been the PolicyViz podcast. Thanks so much for listening.
A number of people help bring you the PolicyViz podcast. Music is provided by the NRIs. Audio editing is provided by Ken Skaggs. Design and promotion is created with assistance from Sharon Sotsky Remirez. And each episode is transcribed by Jenny Transcription Services. If you’d like to help support the podcast, please share it and review it on iTunes, Stitcher, Spotify, YouTube, or wherever you get your podcasts. The PolicyViz podcast is ad free and supported by listeners. If you’d like to help support the show financially, please visit our PayPal page or our Patreon page at patreon.com/policyviz.
