The following piece of code is from the OpenStreetMapX package. The logic is pretty straightforward to follow. I was just wondering, where is the constant 3.6 from? Does it come from some statistical fittings? Any comments are greatly appreciated.

```
### Transpose distances to times ###
function network_travel_times(m::OpenStreetMapX.MapData, class_speeds::Dict{Int,Float64} = OpenStreetMapX.SPEED_ROADS_URBAN)
@assert length(m.e) == length(m.w.nzval)
indices = [(m.v[i],m.v[j]) for (i,j) in m.e]
w = Array{Float64}(undef,length(m.e))
for i = 1:length(w)
w[i] = 3.6 * (m.w[indices[i]]/class_speeds[m.class[i]])
end
return w
end
```