I'm referring to murder because it's the easiest crime to gather statistics for. It's difficult to hide a body, and in this day and age when every murder victim's phone and online footprint will be pored over by police investigators, it's easy to tell when a victim was trans.
Regarding the research about the rates of domestic violence, I would point out that the research has been conducted by Stonewall, which is a special interest lobby group with a particular focus on trans people. They aren't going out, reaching out to vulnerable women in difficult to reach communities and trying to find out whether the rate of domestic violence in the general (non trans) population is underreported. Stonewall is also very far from neutral in this area so I would always take their research with a large pinch of salt.
As for hate crimes, obviously the rate of hate crimes against trans people is higher than the rate of hate crimes against women, because crimes against women are not classed as hate crimes. So only women who have another particular characteristic (such as being a lesbian or ethnic minority) can be the victim of a hate crime.
I'd also be very cautious about any research about "hate crimes" which are not reported to the police. How can we be sure that the things which are not being reported are actually hate crimes? What is a hate crime in the eyes of a victim is not necessarily a hate crime in the eyes of the law. (Misgendering, for example.)
One thing where I do agree with you is that we need better crime statistics. And that means we need to be recording the sex of both perpetrators and victims, regardless of how either of them identifies.