I'm wary of "official density stats" because they inevitably include stuff like Staten Island or a huge chunk of St. Lawrence or Lake Ontario surface area, while my very rough method (explained above) does not have that flaw.
Your own city has pathetic official density, you should be more sympathetic to my method (and I stand by it - I'd sooner take a few random Google Street Views of various neighborhoods and extrapolate from that than take HRM pop divided by HRM area).