In September 2016, FBI director James Comey announced that the agency would begin building a national database to track killings by police. To date, federal records have been woefully incomplete, relying only on voluntary self-reporting from police departments. The Guardian US has been responding to this huge data gap since June 2015, when it launched a project with the ambition of systematically cataloging every confirmed death in the US that occurred during contact with police. Relying heavily on tips from an engaged public, including activists, families of the dead and readers tracking local news, a team of journalists and web designers built an unprecedented searchable database of these deaths alongside original reporting. The NYPD has since announced a program to track deaths systematically, and national attention brought to this issue by Black Lives Matter organizers continues unabated. We spoke to senior reporters Jon Swaine and Oliver Laughland about the ongoing project known as “The Counted.”

CPCJ: Were either of you involved in the initial decision to set up the database?

Jon Swaine: We both were. Katharine Viner used to be the US editor. It was originally her idea after Ferguson, where I had been reporting. She suggested that Oliver and I work together to record all the deaths, then it was left to us to work out how.

CPCJ: When you were putting together the methodology, how did you decide which metrics you were going to track?

JS: Some things were quite obvious to us. It was important to know the race because so much of the debate surrounding this issue, after Ferguson, was about whether African Americans were disproportionately impacted by this issue. So the fact that there wasn’t any official data to prove or disprove that, meant we knew that we wanted to track it; in fact, that was a key reason for doing the project. Other things were basic—like age, sex and location of where these things happened. We experimented with collecting other information, like whether the person was reported to be drunk or on drugs, or whether there were mental health issues, but we found that it was just hard to get that data consistently, so we kept it to ourselves and we used it for research. The race of the police officer involved is another example; the decision of the authorities to release that information is so variable. We tried to collect it where we could, but because it was so partial we didn’t actually include it in the published version.

Oliver Laughland: One of the most difficult things that we are trying to track is the status of each case out of a huge volume of cases—there were over 1,100 cases last year. One of the unique things about our database is that we are keeping track of the ruling in each case, whether the investigation into the death is still ongoing, whether somebody has been charged or whether—as in most cases—there are no charges and the officer has gone back to work. For us, that has probably been the most difficult variable to keep tabs on because it means constantly going back to each case. When you’ve got such a huge volume of them, trying to keep that up is very difficult, but it is definitely worthwhile.

CPCJ: What was the decision-making around the thematic, in-depth stories that you’ve pulled from the data, such as the focus on vehicles or the focus on Tasers? Can you tell us about how you decide which of those angles you want to cover?

JS: We had been doing the project for a couple of months and a few things leaped out. We noticed that there continued to be cases around the country where people were accused of using their vehicles as weapons against police officers and were being shot down because of it. We were looking at deaths following Taser use, whereas other media and people tracking this issue were not, so we decided to make that a key part of our reporting.

OL: There were three stages. The first was when we built the database and launched, but we also wanted to find cases that had been overlooked by national media. That was the initial reporting phase, and then we went into longer reads. In the end, we had the idea of trying to find the place where there was the highest rate of officer-involved shootings, and this place—Kern County, California—had come up repeatedly when we were putting information into the database. The frequency of deaths in that place was quite striking, so toward the end of the project, we went and spoke to the interactives team who did some analysis and found out that Kern County was the place that had the highest rate of officer-involved deaths that year. We wanted to go to that place and do an extended piece of reporting, which became the five-part series “The County.”

CPCJ: Do you have any thoughts about what elements turn an event into a national concern, in contrast to events in very similar circumstances that remain unknown?

JS: One is obviously video. If there is shocking video footage that is posted to the web, then that usually that becomes a case in the public eye. Timing also matters a lot. The deaths of Philando Castile and Alton Sterling happened very close together, and that built up a sense of controversy around this issue again. There doesn’t tend to be that much public sympathy if the victim was committing a serious crime when it happened, even if the shooting or death themselves were quite shocking. Then of course, there’s the element of unrest. If there are protests, media sensation often follows that. Sometimes that’s not true—it is variable—but in William Chapman’s case, for example, there wasn’t any video, there wasn’t any unrest. I think there are some areas of the country where the media are slower to be shocked, especially in the South—and this was in southern Virginia, right on the border with North Carolina. There’s less of a sense of outrage than there is in some parts of the North such as in places like Minnesota, New Jersey, New York, D.C. All those ingredients are in a kind of stew that sometimes causes outrage, and sometimes doesn’t.

CPCJ: Could you talk about why you thought it was important to track other categories of police-involved deaths beyond shootings?

OL: For me it comes down to a human interpretation of the issue. You think of deaths like those of Freddie Gray and Eric Garner, which have become synonymous with the overuse of force and the Black Lives Matter movement. We thought we had to include cases like those in a database trying to shed light on an issue as important as this. It’s much more difficult, and requires a lot of effort not only from us, but from the other people who are involved in this project—for instance, having to find medical examiners’ reports to figure out whether the death meets the criteria of the database. It’s worth it because these deaths are part of the issue, and need to be counted.

CPCJ: How much of your material comes from which sources, proportionally?

JS: The most common scenario is that a death has been reported in some way on a local news site, and a reader has noticed it and told us. We get emails through the website, we have a Facebook message inbox and people reach out to us on Twitter. There are some cases that were never reported, and we find them much later. Texas and California, for example, have begun publishing lists of all such deaths in their states, and when these lists finally come out, we might find two or three there that were never reported by anyone.

OL: The database is a starting point. In some of the long reads, we were looking at dozens and dozens of cases to try and get more information on medical examiner findings, eyewitness accounts of what happened, particular departments’ use-of-force policies and countless other things. The individual cases were often just the starting point for information we were trying to get.

CPCJ: It is obviously a resource-intensive project. Would you characterize this as a kind of higher risk/higher reward model for outlets like The Guardian if you compare it to more standard beat reporting?

JS: I think it is. We were fairly confident that it would be a worthwhile project. We didn’t know if it would be a “success,” or win any awards, or get much attention—so there is a risk. We did devote a fair amount of resources to it. The project takes a few reporters and a lot of time from our interactive team, our video team and social media team, who still work on it every day. It could have been a complete flop. Luckily it wasn’t, but for a middle-sized outlet like ours in the US, you can only pick a few of those things a year. So it does require a lot of thinking about how you want to proceed and whether it is worth it.

OL: In the moment that we conceived the idea, audience interaction was really central to building a community of loyal readers who submit hits, and that happened. “The Counted” has become this extraordinary resource, not just for us, but for other people who are very passionate about the issue. One of the things I’m proudest of is that the database becomes one of the most read things on our site for days, if not weeks, after high-profile cases like the deaths of Philando Castile and Alton Sterling. People constantly come back to it and use it for their own reporting or their own understanding of the issue. That has been hugely rewarding.

CPCJ: When it comes to that group of people who are generating leads for you, do you make a categorical distinction between people who are advocates and organizers and people who are “concerned citizens”?

JS: There’s a spectrum. Some are purely interested in the data and want this to be properly monitored and recorded, some are hardcore activists against what they perceive as excessive use of force by police, and some are protesters who have been in Ferguson or Baltimore or other places. Some of our readers are our critics who are more supportive of police actions and think police have got a bad deal from the coverage from us and other places. There is a wide spectrum, and we are obviously careful not to allow either end of that spectrum to unduly influence our coverage.

CPCJ: In your Reddit AMA, you were saying that readers have vastly different interpretations of the issue. Some attribute it to individual racist police, some to systemic racist profiling and some to excessive criminality in the black community, etc. So you have a lot of conflicting narratives out there. What are your thoughts on the challenges of reporting on these complex sociological questions within the constraints of the kind of journalism that you do?

JS: It’s challenging not to fall into groupthink. The obvious challenge is to keep a clear head and keep reporting what you’ve actually found and to be honest with the data, rather than to let your preconceived notions set the story. I guess that is true for most beats. If you are an environmental reporter or a gender reporter or a political reporter, the main challenge of your job is to be fair, accurate and to keep yourself honest. When you’re covering an area such as this, where there’s a lot of angry and passionate input from both sides, you have to tread the truthful ground. That doesn’t necessarily mean right down the middle, but it’s crucial to be accurate, independent and truthful.

OL: One of the things I was surprised by was how many different beats this issue covers. We have had to get into medical science and mental health reporting. When you have a tool like this database that collects information on so many different data points, you can’t get hung up on one particular sphere. That is what we were trying to do all the way through the project to try and address some of those questions that came up after Ferguson but also to try to expand the issue and look at it in different ways, because that is what the data was telling us to do.

CPCJ: Roughly what percentage of your workload now would be devoted to the project? In the beginning it must have been all-encompassing.

JS: We have cut down on our input, though we are still involved. We luckily have a brilliant colleague, Ciara McCarthy, who largely takes care of the day-to-day maintenance of the database. Last year we did more intense investigation, but now with the election, everyone had to be doing different things.

OL: When there are cases that come up, the information that we have in our database becomes relevant again, so we can go back and get more up-to-date statistics on the number of people killed this year or what the racial dynamics of that are. It’s a constantly growing, constantly updated point of information that you can keep referring back to.