I can't speak about Nantes which I haven't visited for decades but I think it is wrong to count bikes lanes as a single parameter.
Dutch urbanists have found that bike lanes are very important when streets used by cars are unaltered. Once you implement a lot of traffic calming features and cars never reach 30kph comfortably, bike lanes aren't that important and then streets can be shared across all users.
> it is wrong to count bikes lanes as a single parameter.
Of course it is. See % of safe km as a minimal condition, not as perfection.
Note that bike lane ! = safe km. You're right. There are bad bike lanes and good peaceful streets. See other comment for algorithm.
20 km/h streets are counted.