How We Made “Money as a Weapons System”

A Q+A with ProPublica about making an app of goats and sweaters

ProPublica’s recent app, Money as a Weapons System, is a fascinating choose-your-own-adventure journey through the war in Afghanistan. It serves up records about how commanders in Afghanistan spent a $2 billion pot with few strings attached, within a beautifully designed app that’s inextricable from the story itself. It’s addictively browsable and full of kaleidoscopic ways to read and understand the quirky and heartrending stories behind the money. It’s also a solid reminder that a heap of dirty data can clean up real nice. We gave the ProPublica team a call to ask how they made it happen. (We’ve lightly edited their responses for clarity and concision.)

Beginnings & the Data

Q. How did this project get started?

Megan McCloskey: I had been looking into Afghanistan reconstruction and all the money that’s been spent there, which is about $109 billion right now. Part of that is this program called the Commanders Emergency Response Program, which is $2 billion worth of cash distributed to commanders in the field over the course of the war to basically buy hearts and minds. I was told there was an entire database that tracked all this money. We realized that SIGAR, the Special Inspector General for Afghanistan Reconstruction, had gotten a copy of all this data, so we FOIAed for it and got it earlier this year…it was incredibly dirty data that the amazing wizards at ProPublica tackled and turned into clean, useful data.

Q. What was it like working with this data set?

McCloskey: I had covered both Iraq and Afghanistan and had seen this money in play out there in both of these countries. But still I was not expecting the variety, the sheer oddness of it. There were things like, “We’re going to build a girls’ school,” which is something I’d seen before and would have expected. But then there were things like, “I’m going to buy 200 volleyballs and 500 pairs of shoes for the Olympic director” or “We’re going try to do a Mother’s Day celebration,” which is a very odd choice, culturally. You get the sense that it’s one guy or a couple of guys sitting there brainstorming, “What are we going to do with this money?”

Mike Tigas: The data set consists of records that were hand-entered by military commanders while out in the field, so a lot of this stuff is full of misspellings, there are many ways to transliterate place names in Afghanistan…

Ingrid Burrington: I think my favorite mistake is that there was one catch-all category called “other humanitarian projects.” Over a dozen were spelled with two Ms in “humanitarian.” I suspect it was because there were no drop-downs for categories and someone had to hand-enter that every time… Of all the weird entries we found, my favorite was one where the person entering the data started complaining about how crappy the database was.

Design & Code

Q. How did you decide on the “choose your own adventure”-style dropdown headline and overall user interface?

Tigas: Because [the data set] was so messy, we avoided doing aggregates. Rather than saying, “we spent this much on goats in total” we pointed to things that we thought were interesting. That led us to this pile of stuff—this “guided tour” of different categories that we picked out of it.

McCloskey: One of the things I’d been talking about from the beginning with this project on wasteful spending and contracts in Afghanistan is that people just don’t pay attention to stories like these. There’s an eye-rolling effect, we just expect our government not to spend our money well. It was important to me to make it so people would actually pay attention and want to dive into this data. I have a lot of expertise on the military after covering them for a decade, and even I was going through these entries, thinking, this is dense, hard reading. How do we make it accessible? I wanted people to feel like they could play a little bit. As we were going through the data in the news room, someone would say, “Did you see this?” at something funny or odd or surprising, or it would make all of us laugh or run over to the computer and want to read more. We wanted to recreate that experience for the readers using the app… It needed to be easily digestible.

Tigas: As we went on, it seemed less and less like it would read as a chapter-to-chapter story and more like, each of these things could be read in any order.

McCloskey: I kept saying, “We need a button or something where people could pick, and it’d just be all the goats.” Our wonderful team made my notion of a button, and those illustrations just hit it so well. They’re very simple but they still draw you in. It was an element of fun that I thought didn’t undermine the seriousness.

Sisi Wei: There was a phrase that our editor, Scott Klein, used a lot when he was describing the vision. He kept saying: just imagine yourself looking through the miscellaneous bin at a record store and trying to dig for treasure, to find something that was really cool or that stood out. Then it’s about helping people realize what they can search for. If you’re coming into this cold and haven’t looked at the data, you don’t know where to start. We’ve added a long list of example searches that change every time you search. Eventually you can generate your own ideas for what you want to look for.

Burrington: The title “Money as a Weapons System” came out of a phrase and concept used in the military for this strategy. One of the things to take away from looking at all this data is that money becomes interchangeable for so many things that you wouldn’t even think to put a price on. Not just human life but also, “How much does it cost to build a city? How much does it cost to explain the concept of the rule of law?” And so playing off of “money as X or Y” was where the idea for using the dropdown headline came from.

Q. Tell us more about the somewhat whimsical tone of this project, particularly the illustrations.

charming goat illustration

One of Sarah Way’s illustrations brings the data’s ever-present goats to life

Wei: Before there were line drawings, we actually had a version in which Ingrid had found stock art images. So instead of seeing a line drawing of a volleyball, you would see a generic volleyball photo.

Burrington: The decision to have some kind of illustration came out of realizing that all the reports were based on actual things–so if we’d used photographs, you’d want the photograph to be of that actual thing and not some kind of reference image. Having something that was more representative of the idea behind the thing (abstracting it) seemed like a better approach.

Wei: We went through so many different art phases. After a while, we thought, “What if we go to a toy store and buy little toy goats and little toy horses or whatever and take pictures of those?" That sort of became untenable.

Burrington: There aren’t enough good toys!

Wei: Scott said, “Let’s sketch these things, and see how that looks.” [Illustrator] Sarah Way is also Scott’s wife. Since we weren’t sure if we were going to use illustrations at all, he said, “I could just go home, and my wife and I could make some.” When we saw the first illustrations, we thought, “This looks really good. This makes more sense.” So then we started asking her for batches of them. As soon as we saw them in the app, we knew it was the right decision.

Q. We heard that you created a hack involving Twitter cards, to promote the app. How did that work?

Tigas: We had all these highlight cards, and our social team was thinking about ways we could promote the app. We wanted to try an idea where you could share the app by tweeting one of these cards out, and it would be a little preview image. It’s a little harder than it seems because you can’t upload an image and link to a website from a tweet button. (You can from Twitter itself, but not from a button.) So I created a system where each card has a permalink specific to it. If you link to that card, Twitter will give you a preview image of that card–that highlight–that we picked out. If you click on it, it will redirect you to the home page of our app. So if you are on the app and you hover over any of the cards, you’ll see a Twitter image hover and pop up. And if you tweet one of those, you’ll get a little preview image tweeted, of that card that you just clicked on.

Process & Takeaways

Q. What was your project process and timeline like? How did the tech side and creative side come together?

McCloskey: We got the data at the end of January/beginning of February, and we started cleaning it up. It was a couple weeks of just cleaning, and then diving into the design and accessibility of it. I’d say the whole thing took a couple of months…but if you were going to do it with easier data, I don’t think it would have taken that long. Ryann Grochowski Jones, Ingrid Burrington and Jeff Larson kicked us off by cleaning the data and putting it in something we could look through online. While they were doing that, I was able to go in through their very basic app and start searching. Then I could start doing my writing. Jeff got pulled off for another project, and Mike Tigas came in. Then we were talking about accessibility and making it look fun, and Sisi came in, and she’s a rockstar, and she created that sort of “easy” look.

Wei: In general, we make a lot of apps, and the same people that make the app at ProPublica are the same people who write the stories that go along with those apps. The other half of the time, we work with reporters who are in the newsroom who are very familiar with what we do… This app in particular is special because I don’t think we’ve ever made an app yet where the story and the app were the same unit. There is no separate article page or anything like that. If you want to read anything about the project you have to go to the app. Our team also has full control over the tech we use.

Tigas: We used this thing called ArchieML that some people at the New York Times wrote. We like it because they wrote it to match the sort of ad-hoc markup that folks tend to use while editing. Technically, it’s just another way to define key-value pairs (“headline: Lorem Ipsum”), but it’s got a syntax that’s kind to non-programmers. We’re using their Ruby version and followed their instructions to get this integrated with Google Docs.

In development, the app pulls the narrative copy from a Google Doc that Megan and our editors work on, so they can refresh our dev of the app to see how their changes will look right away–without having us touch the app code, or copy and paste text over or anything like that. (In production, we modified this workflow a bit: we save what is in the Google Doc to a flat file, so that the copy is frozen to the version we want. Basically, this is just in case something happens to our Google Doc or if Google changes something about the way our app can access the doc. Keeps it fast, too, since every user isn’t causing our app to hit Google constantly.) We’ve never done anything so integrative before.

McCloskey: One of the cool things was that writing was a very interactive experience. My words were popping up in the app as they were still developing it, which was so helpful. At first, I started writing a traditional story with an anecdotal lead. Then we saw it in the app and thought, “That’s not going to work.” …We were able to move faster because we had this interactive editing.

the google doc and the code

ArchieML + Google Docs = friendly process

Q. What do you hope people take away from this project?

Wei: The concept of war is centered around fighting, but this app is asking, “what are we using money for, instead of fighting?” When they use the phrase “winning hearts and minds,” what did these commanders think was going to accomplish that?

Burrington: A lot of ProPublica’s projects are hard investigative journalism that tend to have a clear kind of villain, or a point they’re trying to get to. So much of [this app] is about how hard it is to actually know what happened–and where and who is really to blame–for a lot of things that happen in war. That’s one of the things that I hope people take away from it: all the complicated human parts that any kind of conflict entails.

McCloskey: I hope they take away the big themes, the chapters, that this program has really spiraled into something much bigger than it was ever meant to be and that has led to some poor outcomes. The very expensive programs don’t seem to go very well. Also a sense of how weird our wars are these days. Fighting counterinsurgencies has really changed the dynamics of who our troops are–I have a line in there about how they’re “insurance adjusters and health advisors” and all these different roles that you wouldn’t expect. We have 20 year-old kids doing somewhat absurd projects.

I hope people get a sense of a little bit more of what it’s like on the ground at war. And from a journalist’s perspective, I hope people see that you can do something that’s interesting and fun and has a lighter tone even if it is a more serious subject, and hopefully it brings people in instead of just tuning people out. I was so terrified this was going to be wonky. I didn’t want just people in D.C. to read this.




Current page