Homicide Watch: An Interview

The central idea of Homicide Watch is radically egalitarian: “Mark every death. Remember every victim. Follow every case.” How did you get here?

Laura: When I first moved to DC, an editor told me that most homicides here were “drug deals gone bad.” I didn’t believe him, but I didn’t have any information to back up my gut feeling that there was more to these cases. But that conversation got me thinking about how we make judgement calls on what to cover and how, particularly when it comes to homicide. My hunch was that the way homicides were covered in DC hadn’t changed much since the drug wars of the early 1990s, and that meant that the assumptions newsrooms were making about which homicides were newsworthy were wrong. So I started from scratch.

I said the goal of Homicide Watch DC would be to cover every homicide in the District. I wanted to know what “normal” was and what “newsworthy” was, but the only way to get at that was to look at every case. Now that the site has a foundation, I’ve come to realize there are many, many more reasons to cover every homicide. To start with, my assumption was correct, only a fraction of the cases are “drug deals gone bad.” But creating a platform and method for covering every case helps many, many more stories surface than a reporter could anticipate by reporting only the “newsworthy” cases. And the site has become a community resource; people come back to the victims’ and suspects’ pages again and again to revisit, to remember, to memorialize, and to connect with others using the pages in the same way.

Then there’s the final reason: covering every homicide allows me to collect data and build investigative projects that would take ten times as much effort as if I had started from scratch every time I wanted to know homicide comparisons by month, or location, or case outcome, etc. Basically, I collect information so I don’t have to spend time FOIAing when I want to do something like this.

On Homicide Watch, you’re sort of exploding what we think of as a “story” into component parts that you can reassemble and update and repackage to follow a developing case over the long term—even after legal proceedings are complete. Is the “news story” still a relevant artifact? And if it is, how will it need to adapt to stay that way?

Laura: Sure it’s relevant. I write news stories almost every day, and most of those stories are plain-old AP-style inverted pyramids. In fact, when you look at the home page of Homicide Watch, what you see is a bunch of news stories in reverse chronological order. But the news story is only one component of publishing and that’s where many news sites are getting it wrong.

To maintain the value of that news story online it has to be placed into its appropriate narrative. At its simplest that means tagging, so a reader interested in the transit board, for example, can follow a transit board tag and see how a story about redrawing bus lines has developed over time.

Homicide Watch takes that one step further by building pages for every victim and every suspect. If a reader is interested in following one case—say, a shooting that happened a block from their house—I want them to be able to find everything I’ve published on that case and I want them to see what the most recent news is. It should be easy for readers to find all the information available, in other words, the narrative from start to…

The other thing is, I look at the information that I’m gathering in the course of my reporting and I try to make as much use of it as possible. That’s where the database comes in. If I know every victim’s age, that information belongs in a database that I can query, in addition to the text of a story. And as long as I have the database, I should make it available to the public, too, because they’re likely interested in knowing how individual cases fit into the larger picture. On Homicide Watch, users can sort by age, gender, and race of victim and suspect and they can query which cases have ended in a conviction or acquittal. We’re always trying to find new ways to open our data up to the public.

This all goes to say that the traditional story is one blip in the larger narrative, and that news sites, in my opinion, better serve readers by making that narrative accessible.

Can Homicide Watch be a model for a new kind of reporting? Where can it be most useful?

Laura: Homicide Watch is data-driven beat reporting built on a framework. Which is to say, a software platform designed to guide reporters through the steps they need to take to successfully cover a story. I certainly think that that concept can be applied to lots of other journalism endeavors, but primarily in beat reporting where reporters can anticipate what sort of information they are dealing with. For education, for example, you might build your platform around schools. For public health, you might build it around hospitals, or illnesses. Basically taking note of any pivot point in your reporting and exploiting that.

This seems like a pretty thorough re-envisioning of the reporter’s job. How does your mindset as a reporter (or a publisher) have to change—or not—to work in a framework like Homicide Watch?

Chris: We think of the site as more of a comprehensive resource than simply a collection of stories. That’s probably the biggest shift from traditional reporting. We file tiny updates sometimes—we’ve identified this victim, this suspect’s hearing was delayed—but those don’t just disappear. Everything we write about a victim or suspect ends up on a victim or suspect page, so we know it’ll be findable later. We try not to throw any information away.

The other side of this is that individual stories are really quick to produce. An initial report on a homicide tends to be really short, with just a map and the few details we know. We can add information later (usually in follow-up posts) and we know it will all flow into this growing resource.

Unlike a print newspaper, the web theoretically offers unlimited space for pursuing stories, though it hasn’t granted us unlimited time. Did this project need the network—both in the sense of technology and of community—to happen?

Chris: Some of what makes the site work is really fundamental to the web: we can link back to profile pages instead of writing a bunch of background in every story. We can see what our users are searching for, and we can organize information the way people are looking for it. Also, we can build a database application that builds our maps for us.

Homicide Watch is sometimes used as an example of “data-driven” journalism, but I have to say, it doesn’t feel like a database or the kind of analytical, number-crunching project I associate with that term. It’s a very human-centric site. How does data enable what you do as a reporter? And conversely, how do your editorial choices shape your use of data?

Chris: In a lot of conversations about data in journalism, “data” is synonymous with “numbers.” We do plenty of numbers stories, but when we talk about data, we’re talking about organizing the information a reporter gathers into a regular and recurring structure. There are patterns in beat reporting, and we can build software that knows what we’re usually looking for, and that will tell you what you’re missing.

For example, every murder victim has a name, age, race, gender, place of death (at the scene or at the hospital), cause of death (shooting, stabbing), and so on. We know when suspects are arrested, for which case, and how those cases end. That’s all data.

What we’ve done that’s different, I think, is woven the data-centric parts of the site into the narrative parts. There’s really no separation. If the site works as it should, a user can land on a story and immediately see which victims and suspects are involved, understand the backstory, and where the case stands. Users shouldn’t have to guess about these things.

In an interview with Alex Howard, you talk about content-centric technology—“news apps” and your own content management system—as tools that allow you to slice and reassemble data to better serve your readers. How important is it to build or customize tools to suit your editorial workflow and aims?

Laura: Having the platform speeds up my workflow by creating a pattern out of my work. So, when a case comes in before I’ve had my coffee in the morning I can pretty much operate on autopilot. Also, information comes in about these cases over time, and sometimes having a new little tidbit, like knowing who the detective is who’s assigned to a case, isn’t worth writing a story about. But I can file it into the database and not lose that information.

The database also tells me where my holes are. About once I week I go through the database and I look for what’s missing: obits, photos, times, anything, and I try to fill in what’s missing. Doing that means I’m always ready to slice and dice the numbers at a moment’s notice. Just recently we had three homicides in one night. As I was writing the story it occurred to me that, combined with deaths earlier in the week, that week might have the most deaths of any this year. I was able to prove that in just minutes and get that story up. I haven’t seen any other reporter with a database robust and agile enough to do that.

The level of engagement you have with people in the community is so strong. Can you talk a little about the ways in which people interact with your work—and with each other—on the site?

Laura: In less than two years we’ve grown from nothing to about 330,000 page views a month. People spend about five and a half minutes on the site. Just short of half of our traffic is to our homepage. Most of it goes to the framework resources: the victims and suspects databases, the map, the photo gallery, the documents library, and victim and suspect pages. About 67 percent of our traffic is organic. People come to the site because they heard something happened down the street, or that “Joe” got killed, or they’re wondering what ever happened with the guy from school who got arrested. They search names and addresses and end up on Homicide Watch.

We see a lot of people helping us with our reporting. They’ll submit names, photos, biographical details, obituaries. Or someone will ask in the comment threads, “Is this Joe who went to such-and-such high school?” or “Is this Joe who worked at such-and-such?” Then people who know the answer will jump in and say yes or no.

I get a lot of emails from people who want their family members’ cold cases added to the site. They’ll email me and say “You’re missing a case, you don’t have my son. His name is this and he was killed in this year.” But I made the decision at launch that we couldn’t cover cases that predated the launch, that it would have been too much work. And it breaks my heart to have to return those emails saying “sorry.” A lot of times those families will end up writing a guest column for the site about their loved one, their experience, and what they want to happen.

The information you post on the site seems to act as a seed, around which all of this other stuff accumulates. Why do these conversations coalesce around Homicide Watch instead of a Facebook page or a traditional news story? And why are they so good, do you think?

Laura: A lot of these conversations were happening on Facebook, certainly, and Twitter, and even obituary sites like Legacy.com. They still are. But Homicide Watch makes those conversations easier because we have all the information gathered in one place. Not only that, but the information is coming from Homicide Watch.

Our readers know that we know how to keep them up-to-date on court proceedings, that we’ll publish news of an arrest as soon as we have it, etc. It takes a lot of the burden off the families to try and keep everyone informed. And it allows other people into the conversations, too.

Not everyone is going to go down to the courthouse to get a copy of the arrest warrant in the case to read what detectives think happened. That’s ok; with Homicide Watch we do that for them so they can stay informed.

Many news organizations still hesitate to moderate their comment sections except in the most extreme cases. How does your comment moderation policy work, and how did it develop?

Laura: When I launched Homicide Watch, I knew that I would have to have a clear method of dealing with comments, because comments on news organization websites about crime stories tend to be really bad. Just mean, hateful, horrid comments.

I wanted Homicide Watch to be a place where people felt that they could really discuss the impacts of violent crime, where they would feel safe doing so, for it to be a place where people wanted to go to have those discussions. So I knew from the beginning that I would have to moderate comments. I also wanted people to know how I was moderating, what factors I was considering, and what was and was not acceptable behavior.

I looked at a lot of other comment policies while I crafted ours and I think what we have works. I always start from the assumption that a comment should be published…a comment has to give me a reason to not publish it, instead of vice versa.

Sometimes I make mistakes. Moderating comments means making tough judgement calls sometimes late at night, or first thing in the morning. I also try to interact with people in the comment threads because I want to show them that I’m reading what they say and that there are real people here. It’s about being active and present and encouraging an active and present community.

What’s the hardest thing about your work at Homicide Watch?

Laura: Oh, I could give a different answer for every day of the week! Day-to-day, moderating comments can be really tricky. It’s a tough line with some comments because you want people to be able to contribute to the community conversation, but there’s certain words, and sentiments (like threats) that I just won’t publish. I try to respect the different ways people express their emotions, because certainly we’re dealing with an emotional topic, and I try to be fair to the commenter, the person they’re writing about or responding to, and the community reading the site more generally. It’s incredibly hard, but very valuable.

Beyond that, most of my grief comes from trying to get recognition for the site. While the DC community has really fallen in love with the site, it’s taken a lot of effort to get those in the journalism community to see this as journalism…not just aggregation (which it’s largely not) or a blog (which it is, in part, but it’s really so much more). And we haven’t been able to get anyone in DC to see this as journalism worth funding.

Where are you headed next?

Laura: I’m thrilled to be a part of the 2013 Nieman class at Harvard as an inaugural Nieman-Berkman fellow. I’ll be studying digital criminal justice reporting, and probably doing some work on how the concepts of Homicide Watch can apply to other types of reporting.

No one in DC has agreed to take on Homicide Watch DC, so the future of the DC site is really up in the air. Without a news organization taking the lead on it, Chris and I are hoping to turn the project into a student reporting lab, using the platform to teach young journalists daily beat reporting, investigative and data journalism, and community management. In practice that means students will be responsible for the daily operations of Homicide Watch DC (everything from breaking news to investigative packages to comments moderation and community development), as well as for contributing to a blog about their experiences. I’m interested in seeing if students can learn solid journalism skills by reporting with the platform. We’ll also be looking at what the challenges and rewards are for young journalists on the beat and what training is necessary to build strong digital criminal justice reporters. We need about $40,000 to get this project off the ground, keep Homicide Watch DC alive, and pay the students involved for their hard work.

About Laura Amico

Laura Amico is founder and editor of Homicide Watch, an innovative platform for data-driven coverage of violent crime. She’s a 2012-2013 Nieman-Berkman fellow in journalism innovation at Harvard. Laura began her career as an education reporter at the Register-Pajaronian in Watsonville, CA. Later, she joined the Press Democrat’s newsroom in Santa Rosa, CA. as their crime reporter. She’s a New York Times Chairman’s Award winner, a Knight News Entrepreneur Boot Camp alum, and has held fellowships with the Online News Association and the Harry Frank Guggenheim Symposium on Crime in America. Laura is a board member of Criminal Justice Journalists, writes occasionally for The Crime Report, and blogs about beat reporting and entrepreneurial journalism at One Reporter’s Notebook.

About Chris Amico

Chris Amico is a journalist and web developer with experience in local newspapers, national news organizations, and media start-ups. He runs the technology side of Homicide Watch and helps out with editing and data-driven storytelling. Beyond Homicide Watch, his award-winning projects include Patchwork Nation, a local-national collaboration between PBS NewsHour and the Christian Science Monitor, as well as Gulf Oil Leak Meter, a simple JavaScript widget that kept a running tally of how much oil spilled into the Gulf of Mexico after BP’s Deepwater Horizon sank. The NewsHour’s oil spill coverage was nominated for an Emmy for innovative storytelling. Since Feb. 2011 Chris has worked as an application developer for NPR’s StateImpact project.

About Homicide Watch

Homicide Watch DC is part of a growing network of Homicide Watch sites across the United States. Homicide Watch DC was recognized as a notable entry in the 2011 Knight-Batten Awards for Innovation in Journalism, was named an “Open Gov Champion” by the Sunlight Foundation, and has been covered by publications such as Nieman Lab, Columbia Journalism Review, and The Atlantic.