This folder contains the data and codes used to estimate the multinomial logit model results presented in Table 2 and 3 in "Teaching Freight Mode Choice Models New Tricks Using Interpretable Machine Learning Methods".
'cfs_2017.csv' is the 2017 Commodity Flow Survey (CFS) public use file.
'CFS2017_Austin_dataprep_for_biogeme.ipynb' includes data cleaning and variable generation steps to prepare the dataset for biogeme.
'CFS2017_biogeme_MNL_Austin_MLpaper.ipynb' specifies both the baseline and advanced MNL models and generates estimation results using Biogeme package.