My experiments thus far have had somewhat expected results, residential free WiFi use is heavier than commercial free WiFi use.
Just for background, I have four commercial locations providing free WiFi and my own residence filling in for residential WiFi. In the commercial locations, I have a retail store, a public space, a doctor's office, and a restaurant.
What I have observed is that in the commercial hotspots; people mostly connect with their mobile devices, and they only connect for short periods. This is good news for businesses considering setting up hotspots for their clients. It means that their customers are not going to linger too long and bog down their network with huge downloads.
On the residential side, I get a good share of mobile devices; but, the really heavy downloads are the result of laptops and possibly desktops with wireless access. These connections are more bandwidth intensive and persistent.
What This Means for Community Network Planning
Residential community networks need a high concentration of bandwidth donors to spread the load of those who are connecting and REALLY using the connection. I knew that throttling my free connection would come in handy.
Ideally, it would be beneficial to tap into the commercial links to feed into residential areas. The reason for this is that businesses primarily use their bandwidth during the day. In the evenings, their Internet connections are largely unused; my data supports this. Residential areas use Internet in the evenings, mostly. Daytime use is minimal.
The problem is, commercial and residential areas are separate, which makes piping Internet connections back and forth problematic.
I don't have a solution; these are just observations.