Scripts to load better data and run network aggregations

[smmaurer]

This PR updates the legacy scripts datasources.py and models.py to load better base data into Orca, including a driving network. The new data can be used to build template-based model steps.

To load the data, run this code:

import os; os.chdir('../')
import orca
from scripts import datasources, models

That will give you:

• parcels and buildings from MTC
• units, households, and persons synthesized by @mxndrwgrdnr
• a tertiary driving network generated by @gboeing

Data download instructions: data/README.md

To generate a table of network aggregations, run this code:

orca.run(['initialize_network', 'network_aggregations'])

Here's a demonstration notebook: Explore-data.ipynb

There is no template yet for the network aggregations. They're defined in configs/network_aggregations.yaml; feel free to edit.

cc @waddell