A city wants to cut traffic deaths in half within ten years. How would you approach it?

A few things to pin down. How many deaths are we talking about a year, and does the city control its own streets, speed limits and enforcement? Is there a budget? I'd start with who is dying, where and how. That comes before any solutions. I'd map every death from the last five years by victim, location, time and cause. Then I'd group the options into three kinds. One is the streets themselves. That's design, like crossings, lanes and lighting. Two is how people drive. That's enforcement and education campaigns. Three is what happens after a crash. Faster emergency response and trauma care save lives too. For each option I'd estimate lives saved per dollar and how long it takes to work. I'd begin with the crash data, since it decides where to focus. Does the city have it?

A few things to pin down. How many deaths are we talking about a year, and does the city control its own streets, speed limits and enforcement? Is there a budget? I'll assume about 200 deaths a year, so the target is 100 fewer, and that the city controls most streets but not the state highways. I'd start with who is dying, where and how. That comes before any solutions. I'd map every death from the last five years by victim, location, time and cause. I'd expect the deaths to be concentrated. Pedestrians and cyclists are probably a big share, a small number of wide, fast roads account for a lot of it, and nights and speed show up again and again. Then I'd group the options into three kinds. One is the streets themselves. That's design, like crossings, lanes and lighting. Protected crossings and narrower lanes on the worst corridors work without anyone having to behave better. Two is how people drive. That's enforcement and education campaigns. Speed cameras and drunk driving checks fit here, and I'd expect campaigns alone to do the least. Three is what happens after a crash. Faster emergency response and trauma care save lives too. For each option I'd estimate lives saved per dollar and how long it takes to work. Then I'd stack them up until they reach 100 lives a year. That also tells the city whether the goal is realistic on its budget. My hypothesis is that redesigning the 20 or so most dangerous corridors and managing speed gets most of the way there. I'd begin with the crash map. Does the city have it by location?