I thought it consisted of a fixed amount, so maybe an error from my part, but after testing with about 20 different networks the scaling factor hovers around 0.32. In other words, all of the measurements from the API endpoint need to be reduced by about 70% (multiplied by 0.32) before matching the "real" values found in the dashboard.
Thought I could generate a regression that predicts the value of the scaling factor based on specific parameters or characteristics of the network, but to be honest, I feel like Im over complicating things, and cant think of possible valid parameters
EDIT: Could it be that the dashboard is displaying average bandwidth usage, vs the endpoint showing peak bandwidth usage? The site does mention "total traffic consumption"