Data of: a global time series of traffic volumes on extra-urban roads

Yet, data on extra-urban traffic volumes is lacking in many regions. To address this data gap, we produced a global time-series of traffic volumes (Annual Average Daily Traffic; AADT) on all extra-urban highways, primary roads, and secondary roads for the years 1975, 1990, 2000 and 2015.

We constructed time series of road networks from existing global datasets on roads, population density, and socio-economic indicators, and combined these with a large collection of empirical AADT data from all continents except Antarctica. We used quantile regression forests to predict the median and 5 % and 95 % prediction intervals of AADT on each road section. The validation accuracy of the model was high (pseudo-R2 = 0.7407) and AADT predictions from 1975 were also accurate.