Mission accomplished
The mission was to train a traffic prediction model and understand what factors create uncertainty and possibly increase costs for the construction company. We can conclude that a large portion of the potential $20,000/year in fines can be attributed to the is_holiday factor. Therefore, the construction company should rethink working holidays. There are only five or six holidays between May and October, and they could cost more because of the fines than working on a few Sundays instead. Of course, these conclusions are for the chosen model – which we can compare with other, better models. With this caveat, the mission was successful, but there's still a lot of room for improvement.
For instance, one thing that could be covered in further depth is the true impact of temp and rain_1h, and features for snow. Our prediction approximation method precluded Sobol from testing the effect of extreme weather events. If we modified the model to train on aggregated...
