I think it would be interesting to see not just the average but also the distribution of walking speeds. I posit the effect might just be “number of people walking to/from work or arrival time sensitive activity.”, with “average age” and “adoption of mass transit” as significant factors.

(Mass transit because you’re more likely to see commuters! Outside big cities you will measure fewer people walking to work, because they drove).

So, strongly correlated with economic activity and cultural factors around timeliness, but no deep psychological explanation required.

I walk fast in the city when I’m going to/from work or trying to get somewhere on my break.

I don’t walk as fast on the weekends, and tourists in the city certainly seem to be in no particular hurry (at least, it seems that way when stuck behind them clogging up the footpath).