A Northwestern University joint initiative of Medill School of Journalism, Media, Integrated Marketing Communications and the Robert R. McCormick School of Engineering & Applied Science. Northwestern University joint initiative of Medill & McCormick School of Engineering.

Chase Davis on data-driven decision making for news projects

Using the theory ‘Hire humans. Not skills. Not roles.’ as inspiration, the Lab’s profiles are Q&As with highly-impressive makers and strategists from media and its fringes, each with unique perspectives on journalism, publishing and communications technology. We’re talking to smart people who are shaping the future of media. Not all of them work in a newsroom, not all are big names, not all have fancy titles, but each is a bright person with something to say. Catch up and/or follow the series here.

If a model exists for the type of young journalist everyone’s talking about and pining for these days, Chase Davis is probably it. He graduated from the University of Missouri in 2006 with a journalism degree and some solid reporting chops thanks to time at the Boston Globe, Milwaukee Journal Sentinel and a few other name-brand newspapers. But then he went and made a name for himself by applying machine learning, natural language processing, and statistics to his work. Check out his Five Algorithms in Five Minutes talk at this year’s NICAR.

He served as director of technology at the Center for Investigative Reporting, helped to build and launch the Texas Tribune, and recently took a job as assistant editor for interactive news at The New York Times.

Basically, he’s a guy worth paying attention to.

Chase Davis

Chase Davis

Ryan: Tell us a little about yourself and your new role at The New York Times.

Chase: I’m an assistant editor on the interactive news desk, where I was hired primarily to work on projects related to politics. What that means in practice is that I get to work with brilliant people like Derek Willis and Jake Harris to figure out A.) How to expand and improve our interactive storytelling efforts in advance of the 2014 and 2016 elections; and B.) How we can use machine learning, natural language processing and other data science techniques to open up new opportunities in our politics coverage.

My background is very much in the “data” part of data journalism, so I’m particularly excited about No. 2. If you haven’t read The Victory Lab, by Sasha Issenberg, I’d highly recommend it. Among other things, it shows how far ahead political campaigns have gotten in terms of using sophisticated data analysis to drive strategy. We’re working on some very interesting projects that I hope will help even the score.

What excites you most about journalism in 2013?

Chase: Not to sound like a broken record, but I think we’re getting to a point as an industry where we’re starting to understand how to concretely better ourselves through data-driven decision making.

I know the Times is an outlier, but on my first day here — literally, my first day — our Knight-Mozilla OpenNews fellow, Brian Abelson, dropped this mind-blowing analysis showing insanely granular details about how people were interacting with one of the team’s Oscars projects. When everyone finally peeled their jaws off the floor, the overwhelming reaction in the room was “we need more of this.”

It’s not just the Times. Conversations I’ve had with managers in other newsrooms over the last few months have me convinced that people get it: Analytics aren’t just pageviews and uniques, and they’re not an assault on editorial autonomy — they’re an opportunity to understand how our audience interacts with us and choose, using our editorial values and sensibilities, how to respond.

As data becomes more important in newsroom decision-making, it’s also going to foster demand for structures and processes that can exploit it. That’s going to mean being a lot more nimble. For example, if we learn that people click through to multimedia more often when it’s at the top of a story versus the left rail, we’ll need a CMS that will allow us to quickly adjust our layouts accordingly.

To the extent that the industry keeps moving in that direction during 2013, I think we’ll be better off for it.

What has changed since you started working?

Chase: I tell this story a lot, even though I probably shouldn’t. But right after I graduated from college in 2006, I was working for the Houston Chronicle, and I wanted to get my hands on a server we could use to host some simple news apps and run some higher-order analysis jobs that wouldn’t fly on a Windows desktop. That request was wholeheartedly supported by my amazing boss Jacquee Petchel, who was the investigative editor at the time, and we started setting up meetings.

Long story short, about five meetings deep we end up sitting around a long, wooden conference table facing a phalanx of upper management suits, who explained to us, very politely, that such a request was out of the question. There were things like Sarbanes-Oxley compliance to consider, and what did a couple journalists know about managing servers anyway?

I say this not to poke fun at the Chronicle — I know a lot of media companies reacted the same way back then — but to me it shows how far we’ve come regarding the role of technology in the newsroom. Maybe I’m lucky, but it’s been a long time since anyone has told me “no” for technology reasons, and I’ve made some pretty crazy requests.

Used to be, the attitude was “better safe than sorry.” Now it’s more and more becoming the opposite. I think that’s great.

You’ve gone from traditional media outlets (Des Moines Register and Houston Chronicle) to less traditional places like Texas Tribune (where you were a consultant) and Center for Investigative Reporting/California Watch, what are the big differences you found between those places and the traditional news shops?

Chase: The easy answer is usually that startups are speedboats and traditional organizations are steamships, but at least for me it’s been a little more complicated than that.

When I worked at the Center for Investigative Reporting, which despite being the oldest non-profit investigative reporting organization in the country was in full-blown startup mode for the three-plus years I was there, no idea was out of bounds. We built iPhone apps, coloring books, custom animations, and a whole lot more to go with our stories. Basically, if someone brought up an idea in a meeting, and it achieved some critical mass of murmuring and head-nodding, we did it. And that was awesome.

That said, we still counted on our traditional distribution partners to get eyeballs on that content. CIR made a strategic choice to maximize its audience through syndication rather than building readership from scratch, which I think was a great decision. But it also meant some degree of begging, borrowing and stealing to achieve maximum impact. Usually it worked, sometimes it didn’t, and I think that’s an experience shared by a lot of media startups these days — including those that have been much less successful than CIR.

Meanwhile, you’ve got traditional media organizations, where doing something out-of-the ordinary requires black-belt political jujutsu to get the right stakeholders on board and build momentum. It doesn’t happen as often, and a lot of good ideas die along the way, but when the all the resources, brand authority, and promotional capacity of a major news organization align behind an idea, there’s not a whole lot that can beat it in terms of pure impact.

So you pick your poison. Sometimes a vision or a goal is better executed in a startup environment, and sometimes you need the weight of an established brand. The trick is figuring out which situation is the right fit.

What companies or media outlets that we might not know about are doing exciting, innovative media, tech work?

Chase: I often find myself looking outside media for inspiration. That’s not because media companies aren’t doing exciting or innovative things — I’m just a shameless thief who loves to steal ideas from other fields.

On the data front, I follow Kaggle closely. They have a great blog, No Free Hunch, that offers up great insights on subjects like machine learning and statistical modeling. Ditto with the folks at Prismatic, who are doing some insanely cutting-edge work, not only around content aggregation, but even more broadly around design patterns for large-scale data problems. Someday I need to learn Clojure just so I can understand what the hell they’re up to.

In other areas, the team at Umbel is doing very cool work using in-depth audience analytics as a tool to better sell premium advertising, which is an interesting way to fight back against plunging online ad rates. Stamen is absolutely worth following if you don’t already, as is FlowingData if you’re into data viz.

Now that I look at it, that’s a pretty disorganized grab bag of random stuff that happens to feature a lot of my friends. But hey — whaddya do? =)

How can people use technology to create change in the world?

Chase: I don’t believe technology’s ability to change the world is especially unique. Honestly, I think those of us who are technologically inclined are flat lucky that the world happens to value what we do right now. It might not always be that way, but at least for now we’re in the right place at the right time.

For years I’ve been telling students that if all this programming stuff I teach doesn’t resonate with them, that’s fine. I don’t think all journalists, or plumbers, or mayors, need to be crack developers (although having a conversant understanding of code doesn’t hurt). Instead, they should find something that inspires them the same way technology inspires me then go make their mark with that.

That said, I do think there’s a wonderful attitude in the programming and startup communities that the only barrier between a problem and a person’s ability to solve it is knowledge — and that knowledge can be acquired with enough effort. No need for special training, or money, or pedigree, or, frankly, outside help of any kind. It’s a very no-excuses self-reliant worldview that resonates in this weird way with my Midwestern upbringing, and I wish that attitude was more universal.

What makes good content?

Chase: To me, this is a particularly fascinating question when you look at it from a literal perspective.

I read a bunch of general interest magazines — the New York Times Magazine, the New Yorker, the Atlantic, Texas Monthly — on a regular basis. I don’t care what the articles are about; I don’t care who wrote them or whether my friends have recommended them; I just like reading them because I know I’ll come away entertained by a good story and enlightened by some interesting insights.

So to me, that’s good content. But what does that mean at an atomic level? The topic of the story sometimes matters, but I might also be drawn in by the length of the story, the tone, or the writer’s voice; the brand quality of the magazine, which signals that the stories will be high-quality; or even granular elements that I might not even be consciously aware of, like sentence length and flow; the level of the vocabulary; or the way the story is structured. It’s probably all of those things and many, many more that I can’t even begin to understand.

In that sense, I think we’re very much in the early stages of understanding why different types of content are “good” to different people. Now you’ve got these content recommendation engines that focus primarily on subject matter and social signals. I know you like articles about baseball, your friends seem to be reading these, and so we assume you’ll like them too. And that’s great at making our daily information intake more manageable. What it doesn’t do is expose to the deep, latent attributes that I might associate with good content, and which, in turn, might provide some much-appreciated serendipity.

What applications do you have open while you’re working?

Chase: My girlfriend will confirm that I’m the kind of guy who arranges his TV remotes on the coffee table at perfect right angles, so it should come as no surprise that I like to keep my virtual workspace as clean as possible.

Stickies are always open for notes, ideas and to-do lists. I code in vim and SublimeText2, and I usually have iTerm2 open for shell access. In the browser, I always have one tab open for Gmail and another for Pandora (lately: Guy Clark and Townes Van Zandt). Other than that, I close any apps I’m not using and try my best to keep my Desktop clear of old files. If I’m not using it, I have to close it. Yup. I’m that guy.

Follow Chase Davis on Twitter, @chasedavis. Find weekly updates from our profiles series on Fridays.

Like what you see?

Northwestern University Knight Lab advances news media innovation through exploration, experimentation and education. The Lab's free publishing tools help to make information more meaningful and promote quality storytelling on the Internet.

About the author

About Posted on May 17, 2013 Posted by

Ryan Graff


News nerd ecstatic for the future of news. Formerly a Colorado-based reporter and magazine writer. Presently the Lab’s editor, and handler of marketing and outreach.