Your labels aren’t clear. They assume the reader already knows exactly what you’re presenting. Pls fix

No you’re not. You’re combining random data on the y axis to obfuscate the correlation from the original post.

This is the right response. The y-axis data doesn’t make any sense. You may as well throw auto fatalities in there, since we’re just combining random-ass data. Training police more will probably have the same amount of impact on auto fatalities as it will violent crime.

Am I right in saying that you added violent crime committed by non-cops into the mix, and the correlation disappeared? So, doesn’t that just imply that there’s not a correlation between non-cop violence and training? If that’s true, that’s actually a bit surprising, since I’d actually think/hope that the areas with higher violent crime rates would train their cops more, given that they’re more likely to be put into difficult situations… which leads us to a possible reason why there was a strong-ish correlation in the original chart: cops aren’t necessarily trained enough for what they’re going to experience.