A little more than a year ago, my colleagues and I started the StreetCred Police Killings In Context data project. There were already several projects tracking police shootings, including Fatal Encounters, and new projects at The Washington Post and the Guardian. But we felt that they left out crucial context. Why did the police become involved in these incidents in the first place? Was their use of force something that most people would consider reasonable, such as the police response after Mohammad Youssef Abdulazeez opened fire on military personnel at a Chattanooga, Tenn., recruiting station? Or was it unreasonable, as with the shameful killing of Walter Scott?

The most common criticism I received was that it was an attempt to demonstrate that police aren’t the problem. I’m a white Texas cop, so that’s not an unreasonable presumption. But I hope you’ll see that what we found can steer the conversation toward more productive and meaningful reforms.

We began, as others did — with the Fatal Encounters database. Where The Post’s Pulitzer Prize-winning effort focused on police shootings, we focused on all police-related deaths. And where the Guardian included killings by, for example, prison guards, or traffic accidents, or personal disputes between officers, we focused only on official acts (on- or off-duty) by police officers.

Ultimately, we gathered more than 70 contextual data points, from the mundane (middle name) to the specific (if the person was listed as mentally ill, was it a previously diagnosed serious mental illness, or merely suspected?). We eventually identified 153 cases in which unarmed civilians had died after an encounter with police in 2015.

We now had a wealth of contextual data that we believed could add a lot to the ongoing dialogue about police shootings. We think the context for these shootings transcends both race and politics. We have assembled our findings in a book, “In Context: Understanding Police Killings of Unarmed Civilians,” which we believe is instructive for activists, police officers and use-of-force trainers. The context behind these shootings can also help readers understand how police and trainers approach these cases, and the applicable law they use to determine when different levels of force are justified.

We tried to be objective, but true objectivity is, of course, impossible — everyone has inherent bias. Instead, we’ve tried to be as open and transparent as we can. We included more than 500 source-footnotes in the book, as well as an open update system to track when we got something wrong. The data, methodology and schema on which we based our conclusions are all open and available under the Creative Commons Share Alike license. So please feel free to check our work.

From my interpretation of the data, there are some relatively easy changes police and policymakers can make that will go a long way toward improving policing in the United States. Here are what you might call the “low-hanging fruit”:

Police must release more data, and do it sooner

When agencies delay, public suspicion and speculation fill the void. Every time police choose transparency, it’s a deposit in the bank of goodwill. It builds community trust. Las Vegas has shown that a police department can be transparent and forthcoming after a police shooting without disrupting the integrity of the investigation.

Honesty begins with a timely release of the earliest available information. That might include a timeline, recordings of 911 calls, a narrative and any video. Without video or 911 calls, the public has nothing but the word of the police — and as we have seen, police do sometimes lie.

Not always. Not even a lot, according to our research. But consider the Walter Scott shooting, and the apparent lies told by Officer Michael Slager. A high-profile case like that can quickly make citizens lose faith in the system. Sunlight is the best disinfectant.

While the media and activists groups often focus on race, the data show a greater disparity in deadly engagements with drug addicts and the mentally ill.

In 2015, nearly half — 46 percent — of the deaths of unarmed civilians after a police encounter involved someone suffering from serious mental illness, overdosing on drugs, or both. We tend to think of both mental health and drug addiction to be issues at the margins of society. These issues need far more social, legal and policy attention than they’re getting.

One approach gaining popularity is the Memphis Model Crisis Intervention Team (CIT) for first responders. Under the CIT model, police officers are trained how to recognize the symptoms of a mental illness crisis, how to communicate with people experiencing a mental health crisis, and how to intervene without escalating. The model has been widely praised and should continue to be adopted around the country. But it is clear that the problem is bigger than just crisis management — the work needs to start before a mentally ill person goes into crisis in the first place.

We’re also seeing new programs to address that problem, with promising results. A program in the mid-cities between Dallas and Fort Worth, for example, has been successful in training cops on new ways to liaise with the mentally ill. By focusing on wellness checks in addition to crisis management, the area has seen a reduction in acute mental health crisis hospitalizations. Such creative programs stress regular visits and include outreach to friends and family. It’s a really smart way to work; it saves lives and money and arguably makes the community more strongly supportive of all its residents. Many state police academies have also been adding more training in mental health and disability. In my opinion, we need more of these programs, and with an even broader focus.

But even with new training, we must remember that the police should be the last line of defense from mentally ill people who potentially pose a threat to the community. Instead, police are too often the first. The federal government has long passed the buck on mental health to the states — under both Democratic and Republican administrations. State governments have also dropped the ball. Mental illness is a society-wide issue. It can’t just be a police issue. The buck can be passed down, but cops can’t pass the buck back up. And if we are going to insist on asking cops to respond to mental-health crises, the least we can do is give them the proper training and support.

Look at all disability, not just mental illness

When we add physical disability — such as pulmonary emphysema, enlarged heart, hypertensive or atherosclerotic cardiovascular disease, thyroid disorder — to that statistic, the number of unarmed citizens who died after a police encounter suffering from drugs, mental illness and/or physical disability is 52 percent. That, of course, is more than half.

We need to consider as a society the effects our national health-care policies can have on policing. We also need to be make sure police are properly trained to interact with people with disabilities such as deafness or autism. Health care and policing aren’t unrelated issues. Cops alone can’t change this. It should be part of the larger conversation among activists, health-care professionals, journalists and law enforcement leaders. While it’s true that fixing the American health-care system isn’t exactly “low-hanging fruit,” one fix that would go a long way is better coordination among police, health-care providers and advocacy groups for people with disabilities. The Americans With Disabilities Act isn’t a suggestion; it’s a mandate. It’s time our society faces the fact that when it comes to the cops, it cannot be a mandate that is unfunded and unsupported.

We need more police video, but it must be used correctly

Video was available only in 26 percent of the cases in which an unarmed person died after an encounter with police. In two 2015 cases, eyewitness-shot video was crucial in showing that officers had lied to investigators. In two other cases, video exonerated officers. But in all four of these cases, the video was taken by bystanders or surveillance cameras, not by police-operated video.

Here’s another metric first observed by San Jose independent police auditor Walter Katz: Last year there were zero indictments of police officers in cases that did not involve video. In every case in which an officer was indicted, video was available and became an important part of the evidence supporting indictment. That’s an important metric. We need more officer video, in more places, right now. I believe proliferation of body and dash-cam cameras will mostly demonstrate that police usually get it right and perform their jobs with professionalism. But it will also show when we are at our worst. That, too, is important.

I am a strong proponent of body-worn video, but it’s not as easy as merely handing out the cameras. As we’ve already seen, there are significant challenges to developing practical policies and procedures that bring transparency, while still protecting citizen privacy and police officers’ due process. Rank-and-file officers, union representatives (where appropriate) and community leaders and activists should be consulted during the policy-building process so that there are no surprises. Sometimes, policy can be counterintuitive (for a truly surprising read, have a look at the how the Los Angeles Sheriff’s Department handled the thorny issue of whether officers should be permitted to view video before writing their reports). There are also logistical, technological and financial barriers that need hashing out.

Deaths in police custody aren’t always from shootings

Police shootings get most of the media coverage, but in nearly half of the 153 cases in which cops killed an unarmed citizen in 2015, no shots were fired at all. The decedent was killed in another manner.

In most of these cases, autopsy reports found that the deceased had an abnormal reaction or complication after officers deployed tools or techniques with the intent of using of non-deadly force, such as stun guns or chemical spray. Many of these cases include the drug-related, mental-health-related deaths described previously, as well as other incidents in which decedents with chronic health conditions or a physical disability attacked or resisted officers or others.

In 23 of the 83 cases (27 percent) in which there were shootings, officers deployed a Taser prior to shooting. To my colleagues and me, this is a good indication that officers were attempting to subdue suspects who ultimately died by first using non-deadly force. That would be consistent with current best practices.

That conclusion has frustrated some observers, because activists and reformers have pointed to Tasers as weapons that cause too many deaths. Part of the problem is that there’s a lot of confusion in the general public about a weapon’s potential to cause death and whether it should be considered “deadly force” by law, practice and training. Without devolving into a festival of scholarly citations, the law is quite clear about this: “Deadly force” is force that poses a significant risk of death or serious bodily injury. Legal experts have tended to agree that “significant risk” means “more likely than not.”

Some people have indeed died proximate to Taser exposure, but the numbers are very small. Independent, non-Taser-funded research has repeatedly shown that the percentage of people exposed to a Taser who die soon after that exposure is minuscule, and even among those cases, it’s difficult to establish a causal link (see, for example, this Arizona State University study or this NIJ study). There have been millions of Taser exposures, and very few deaths. In definition and in practice, Taser is non-deadly force.

Our data show that in nearly a third of the cases in which officers used deadly force, they first attempted non-deadly force. Yes, there have been awful examples of Taser-proximate injuries suffered when an officer used the weapon not because a suspect was fighting, but to gain control, or out of anger. That is generally a fireable offense (and sometimes a criminal act). But that’s a separate issue from the cases my team studied, in which police officers faced a real threat and deployed a Taser to stop a potentially deadly incident by using non-deadly force.

Consider, if you will, the death of Omar Lopez last year. Lopez had been fighting a passenger on the SEPTA train in Philadelphia when a police officer intervened. Lopez bit and violently fought with the officer. The officer, armed with a .40 Sig Sauer pistol and a baton, continued to struggle and ultimately used his Taser alone. Lopez died, but the autopsy showed his cause of death to be an overdose of PCP. That officer, like officers in many of the incidents we studied, showed tremendous restraint. When we credit officers who first opt for non-deadly force, others are less likely to reach first for their guns. I think reasonable people would agree.

Each of these reforms is individually actionable. None is dependent on the others, and few seem all that divisive or controversial. With a couple of exceptions, the solution for each is relatively inexpensive, relatively uncomplicated and relatively understandable.

Once we can see beyond the “how many” of police-involved killings and look instead to the “why,” we can start to have meaningful conversations about how to reduce them.

Nick Selby is a Texas police detective who specializes in crimes that leverage the Internet, such as child exploitation, fraud and organized retail crime. He is co-author, with Ben Singleton and Ed Flosi, of “In Context: Understanding Police Killings of Unarmed Civilians.”