Automatic generation of high quality data dictionary

[jpn--]

ActivitySim currently is capable of generating a simple data dictionary using the write_data_dictionary component. The result is a listing of table columns in each table, giving name, dtype (int, float, string, etc.) along with the checkpoint where the column was added. No info is included about what the values mean (e.g. what modes are available, what are the units of measurements, etc.).

We would like it to generate a more robust and high quality dictionary output, similar to static dictionaries provided in these examples:

• Oregon DOT: https://github.com/RSGInc/SOABM/wiki/Model-Outputs#individual-trips-file---indivtripdata_iterationcsv
• ARC: https://cdn.atlantaregional.org/wp-content/uploads/abm-data-dictionary-2017.pdf

[guyrousseau]

Following up on Jeff's example above, we updated our ARC ABM data dictionary earlier this year, see https://cdn.atlantaregional.org/wp-content/uploads/abm-data-dictionary-feb-2022-2.pdf