From Curtis's link (paraphrased a bit):
A study of the 2000 school shootings from 1970, 200 of those where accidental discharges, as he noted. No information if anyone was hit by any bullets or killed by these discharges (except one, where he shot himself)
They also note a study from 1999-2018 68 examples of shootings where an school security officer was present (they did not say if they where armed, but I guess we can assume?). Only 2 cases where the security officer gunned down active shooter (no info on if this was before or after they killed / shot anyone).
This is very incomplete data, especially when you consider what are being called "school shootings" (e.g. many of which in the last year where gun fights in parking lots, as shown in a previous posts link, which very much could have been a neighborhood thing). Not well defined at all for what I think the point is here, so less than useful.
How many people where killed in all these discharges / shootings? How many where injured? How many people would have been shot or killed in the cases where the security officer gunned them down (obviously no real way to know that). And as noted, no way to know if shootings where avoided because they knew they where there.
These obviously are meant to speak to the increase danger in simply having weapons on campus, but as noted, they really say nothing of the obvious trade off, which is really what we are talking about here: Yes, putting guns in the hands of security officers increases the danger of those guns being used, but is there a net benefit (e.g. more lives would have been lost / more shot, if they were not there). This data cannot reasonably conclude that from what I can tell.
The "arming teachers" thing also seems to get heavily conflated with the concept of "requiring them to carry" rather then the far more reasonable (and likely I would hope) "let them carry if they want to and they are well trained".
I am sorry to say, but the "stats" this guy stated seem to be more of a way to confirm what he believes than an actual objective analysis of the situation (no need for a shocked face, it's pretty much the norm). A true statistician would never present such data as a way to draw a reasonable conclusion (one actually already stated that above).