Certainly there is some data with respect to how many people are held pretrial, but if you're trying to drill down into public safety, I think it's a different kind of data, particularly if it's about public safety with respect to what happens when certain people are released on bail. Really the data you would need is about whether, upon rearrest, the system was operating as it should, and whether the new arrest is simply an unforeseeable risk, or reflects some sort of problem with the original bail hearing that maybe could have been corrected by way of an appeal, or reflects a gap in the existing laws governing bail.
Certainly I suggest that the existing laws governing bail give prosecutors all the tools they need to oppose bail in serious cases. When people are out on bail and they are picked up and charged with new offences, that does not necessarily mean the existing laws aren't working.
How do we measure that? Again, because I'm a lawyer and not a criminologist or a statistician, I can only guess with respect to how best to measure that, but I think there are people who are in the business of measurement who could probably answer that question better than I can.