In this episode, hosted by Megan Calcote, Brandon Quester and Evan Wyloge of the Arizona Center for Investigative Reporting share their tips for finding and using census data. Wyloge and Quester share their experiences searching through large datasets and what they have found to be more useful to drill down the numbers to your local area. They also share the websites and reports they find to be the most useful for story inspiration. They originally shared these tips during Reynolds Week 2016.
To read the AZCIR’s four-part investigation on hazardous chemicals, click below:
- Poor oversight cited in Texas ammonium nitrate blast mirrored in Arizona
- U.S. oversight of ammonium nitrate insufficient, GAO says
- 2.8 million Arizonans live within vulnerable zones from toxic chemical leaks
- Federal report recommends overhaul of U.S. chemical safety oversight
Transcript
[Intro music]
Megan Calcote: How to Cover Money: Finding and using census data in your stories.
Brandon Quester: The tough thing about Census is you can get so deep on one very narrow topic that you could spend an entire presentation on one thing. And that’s part of the challenge, and it takes a long time. Like being new to going through the Census can be a little daunting and frustrating, because there’s so much in there, and figuring out exactly what you need can be a bit of a challenge.
Calcote: Hello and welcome to the Reynolds Center podcast. We are coming to you from the Donald W. Reynolds National Center for Business Journalism based at the Walter Cronkite School of Journalism and Mass Communication at Arizona State University. I’m Megan Calcote, host of the How to Cover Money podcast. Today we share tips from Arizona Center for Investigative Reporting, Executive Director Brandon Quester and senior reporter Evan Wyloge on ways you can use Census data in your reporting. Brandon and Evan originally shared these tips during Reynolds Week 2016. At the Arizona Center for Investigative Reporting, Brandon and Evan regularly use Census data in their reporting because they’re often interested in seeing how certain demographics are impacted by the stories they cover. One of these projects included researching local environmental protection agency risk management plans for facilities storing hazardous chemicals such as ammonia or chlorine. Using the data they gathered from the EPA and demographic data pulled from the Census Bureau, they calculated that nearly 3 million people live inside potential blast zones in the state of Arizona. You’ll hear them reference some of the work that went into that investigation today, and we’ll include links to this four-part investigation on the show notes for this episode.
Quester: The tough thing about the Census is you can get so deep on one very narrow topic that you could spend an entire presentation on one thing. And that’s that’s part of the challenge, and it takes a long time. Like, being new to going through the Census can be a little daunting and frustrating, because there’s so much in there, and figuring out exactly what you need can be a bit of a challenge.
Evan Wyloge The way that the data is compiled is by geographies, and they have a bunch of different geographies, and they kind of fit together like smaller pieces into larger pieces. And so you can look at the most granular which are going to be your Census Blocks. Census Blocks are, I believe somewhere between 400 and 1000 people. So that’s the smallest unit that you can look at. Then you’ve got those that fit into block groups, and then those that fit into tracks. And then from there, things get a little bit more varied. They start to compile tracks into things that are called places. Place is typically a municipality. Then you’ve got counties. Places sometimes across counties. So then you start to have different ways of looking at these things, and they don’t necessarily all fit into the larger piece. But can attack them in whichever way it suits
you.
Calcote: So if you’re going to work with census data, it’s a good idea spend some time just playing around with the data to see what’s there. Once you become more comfortable with looking at sorting and understanding the available data sets, you can use it in an almost infinite number of ways. Census data can be used as the basis of your story, or it can be used to add context to data you collect from other sources.
Quester: You guys probably use BLS, like Bureau of Labor Statistics, quite a bit. If you’re pulling a data set from BLS, you say, “I wonder if…” or “Does this population have any impact on that,” like, “If this? Then is that happening?” Think about your stories and your data in that way, and think about how the census data can add to the context of your story, or add to some deeper element beyond that kind of surface level, “Oh, here’s some raw numbers on unemployment. “You know, if I want unemployment numbers, let’s look at it at a more detailed and finite breakdown. A majority of the stories, I think that we do, use census data in some way, because we’re often looking at demographics, and we want to know, if I’m looking at blast zones, do they impact this demographic more than they impact that demographic? Or this region more than that region? Or if one industry is failing in one part of the city, is that going to relate to the unemployment rate through BLS? And yada yada yada. So if you think in terms of not just one data set as explaining everything, but starting to look at, I can look at the racial demographics here, the socio-economic here, combined with this data set, you can start to get into some really complex analysis that you can hang an entire story on. That analysis might take you a week or two, but that finding might be your lead. And it’s really cool. Once you start to get familiar with how to look through census, how to look through other data sets, you can really start to see how you can add context to stories, that even a daily story. And even if you go through census reporter, and you pull some basic numbers, all of a sudden, you’ve added depth and context to a daily turning story that otherwise would have just been kind of surface level engineering.
Wyloge: And something else to I think, to point out is that what you can do with census is you not only have just counts for each of these, you’ve got pretty deep demographic data, right? So you can ask the question, what’s the correlation between multiple blast zone coverage and median income? Right? So is it the case that lower income people happen to be covered by nine blast zones and higher income people are covered by one or none?
Calcote: When you’re ready to start digging into census data, Evan recommends censusreporter.org as a great tool to explore the census in a user-friendly way.
Wyloge: They went out and found reporters that were working with census data already, and they asked them, “How can we take census data in like a wholesale manner and put it all together in one place that’s a whole lot easier to interface with than what the Census Bureau does,” right? So for, you know, for people who want something easier, this is a really awesome website.
Quester: It’s a lot more user friendly than not knowing much about the data and trying to navigate their kind of maze of different data sets and record layouts.
Wyloge: Yeah, what you might most immediately recognize is like, how simple this looks, right, compared to the Census Bureau. Okay. So right away you can just see that they’ve got a profile section at the top. You can just type something in there. They’ve got an explore tool where you can drill down into things, and they’ve got just topic breakouts in case you are just really looking for something broad and want to start exploring.
Calcote: If you’re interested in import and export data in your area, Brandon recommends using the census tool USA Trade Online.
Quester: In October of last year, they opened this up for free for anyone to use. And to do that, you need to go on and sign on, but it allows some kind of nice options for you to pull commodity based data on trade, imports, exports, etc. There’s a foreign trade section, which you can usually get to and explore data, but most of that is in kind of an aggregate way. And with the USA Trade Online. Once you sign in, you can get to the screen here, and it shows the different levels of data that you can find at a state level or regional level, both imports and exports globally. And the cool thing about this system is that you can drill in pretty decent. You guys know, NAICS code? You can spot commodity, commodity type, type of import, type of export. It’s collected by month over a period of time, and the current data only goes back to 2011 but I’ve just started kind of exploring it. I haven’t dove too deep into this yet, because it’s the access is new and I was trying to look at some imports into Arizona, particularly through the Nogales port of entry, which is a mainland port of entry for Arizona. But you can do this basically anywhere, and especially those of you that live in a larger port area, you can get some really specific, detailed data on what’s coming in and what’s going out. Which is pretty cool. In the foreign trade area, you can actually define it like, say, I was interested in, “Okay, what’s coming in through Arizona, and how big of an impact has NAFTA had on trade between Arizona and, Mexico, etc.” You can actually drill down into defining only NAFTA open goods, basically, which is pretty cool. If you have a specific port, it can go through that exact port. It’s pretty detailed. And, I mean, you could spend a lot of time kind of drilling down and seeing how far you can go. And really, when looking at this kind of stuff is, I usually start with, “Okay, what kind of story are we working on?” And if I want to look at like I did a story on food imports into the U.S. and I was looking at cantaloupe specifically and specifically from Guatemala. I ended up using FDA data and USDA data on import refusal records, which is a whole different set of data. But you can really drill down. And if you have a premise, or you have a question, I want to answer this question. I want to know, “What is trade doing for cantaloupes over the last four years?” I can say, “Is it increasing? Is it decreasing?” If I see a big decrease in one year, then I’d say, “Okay, what happened here” Was there an import alert? Was there an outbreak of some sort?” Which there was when I, when I was when I was looking at was back in 2011.
Calcote: Once you’ve collected the data you want to work with, Brandon has these recommendations to help you protect your data source from errors, and to help you keep track of the steps you’re taking during your analysis.
Quester: The very first thing you should always do, whenever you download a data set, make a copy, never work off the original data set that you’re working on. So I always have an original folder where it’s the original data set, and then I have a working folder. And then I have different versions of my working data set, because some of the stuff that we’re doing, we’re doing 10, 15, 20, 30, queries that we’re manipulating the data. And if at the end of it, you turn in your story to your editor, and they’re like, “I don’t believe that 44% of the state’s population lives within a blast zone.” You can track that actually, on that, that hazmat story, one of the things our first query that we ran with the shape files, and it was like, oh, 65% of the state’s population lives within a blast zone. I’m like, “That seems way too high.” What we found is that the overlapping boundaries were double counting. And so instead of having one shape file for the whole state, for each of those, we had individual laid on top of each other. And so that was kind of a gut check of 65% doesn’t sound right. That seems way too high. If it was right, we’re like, sweet.
That’s a really good story, you know, but it was wrong. But always make a copy, and then always kind of do a gut check and, and I’m a huge fan, and we do it track everything you do. Kind of call it like a data dictionary. You can you can go back through and anybody can come into your data set, whether it’s an editor, or somebody that is fact checking, and say this is exactly how this analysis was done. And you can repeat it and hopefully come to the same result.
Calcote: And if you want to stay up to date on new releases from the Census Bureau, Evan recommends signing up for their email alerts.
Wyloge: You can sign up with the Census Bureau to get emails, and they have a whole menu. You can say, I want this information and this information only, or you can sign up for the whole thing. And if you do, then you’ll get, like, five emails a day. And so you may not, may not want to go, or maybe try it. Go, sign up for all of them, and then you can sign back in later, after a month, and say, I only like these five emails. But they’ll send you, you know, stuff that’s coming out, new analysis that they’re producing. There’s also an embargoed census release sign up, so you can get data sometimes 3, 4, 5 days, maybe a week, ahead of when they release the stuff publicly. That way, you can get a jump start on a story.
Calcote: Thank you, Brandon and Evan for sharing your expertise with us at Reynolds Week 2016 and thank you listeners for tuning in to another episode of the How to Cover Money podcast. The winners for the 2016 Barlett and Steele Awards were recently announced. Visit businessjournalism.org to see the three business investigations that received recognition this year. If you’re in need of more business journalism training, the Reynolds Center can help. Visit businessjournalism.org to find articles and self-guided training, download our free eBook: Guide to Business Beat Basics, register for our free email course on covering financials, or sign up for our weekly newsletter. The newsletter will keep you up to date on training opportunities from the Reynolds Center year round. If you enjoy the How to Cover Money podcast, make sure to subscribe on iTunes, Stitcher, or Sound Cloud, and while you’re there, leave us a rating or a review to help make the podcast more visible to other business journalists. Support for the How to Cover Money podcast comes from the Donald W. Reynolds National Center for Business Journalism. Join us on our next episode, when we share tips on interpreting financial statements from Tom Contiliano and Philip Drake.
[Outro music]
