Stats can be so misleading.
Here is the abstract summary from the source of the graph you posted:
This paper assesses the claim that sanctuary cities – defined as cities that expressly forbid
city officials or police departments from inquiring into immigration status – are associated
with post-hoc increases in crime. We employ a causal inference matching strategy to compare
similarly situated cities where key variables are the same across the cities except the
sanctuary status of the city. We find no statistically discernible difference in violent crime
rate, rape, or property crime across the cities. Our findings provide evidence that sanctuary
policies have no effect on crime rates, despite narratives to the contrary. The potential benefits
of sanctuary cities, such as better incorporation of the undocumented community and
cooperation with police, thus have little cost for the cities in question in terms of crime.