[Hali87]

I feel like I understand what they're trying to present but I'm not sure how useful that is in isolation. Basically it's "how far from optimal is traffic flow". If you look at the Canadian cities for example, the gradient of Vancouver (frequent, heavy congestion) - Halifax (moderate congestion) - Edmonton - (relatively little congestion) makes sense; you're more likely in Vancouver than in Halifax to be going below speed limit due to other vehicles obstructing you and less likely in Edmonton than in Halifax.

Commute/intra-regional travel times would be functions of distance and congestion along with other factors (speed limits, traffic lights etc). So between two commutes over roughly the same distance, the one with higher congestion would tend to take longer. Between two commutes with similar congestion, the one with longer distance would tend to take longer (all else being equal).

An instance where this metric might be more important is in terms of adding new transportation infrastructure to the existing network. In this case, adding things like transit or bike lanes on existing streets would tend to require fewer compromises in Edmonton than in Halifax, and would require more compromises in Vancouver (using Halifax as a baseline, there's "excess" existing road capacity in Edmonton and a defecit of existing road capacity in Vancouver, based on congestion).

Anecdotally London ON does feel about as congested as Halifax but more spread out over the course of the day. In Halifax there's heavy congestion during peak hours and little congestion the rest of the time. In London it's more like moderate congestion all the time.

Having Kitchener-Waterloo at the bottom makes sense because in addition to having above-average freeway infrastructure the K-W area also has few natural obstacles and multiple focal points.