The story you link literally says:

"the varying ways to count and then interpret police data means it’s worth taking all claims about crime rates rising or falling with a pinch of salt"

They go to great pains to point out that the data is not recorded consistently, so drawing any conclusions from it is fraught.