SANDAG ABM3 Land Use Preparation Tool

Tool for preparing land use and parking inputs for ABM3 from Estimates and Forecasts outputs. Household and person processing is handled by the Population-Sim pipeline. There are three main parts of the tool:

1. Process outputs of the 2022 parking inventory to impute missing data.
2. Estimate regression models to predict the number of free and paid spaces in an MGRA.
3. Create land use inputs for running ABM.

Unless there are changes to the base year network or inventory, only the third step needs to be run.

A dictionary of the settings in config.yaml can be found here.

To easily run without parking preprocessing

1. Gain access to the database RP2025 on the server DGISWSQL22 from GIS.
2. Create a directory to run in.
3. Clone the repo into the directory.
4. Create anaconda environment using the environment.yml file.
5. Open up config_[YEAR].yaml and set run_ABM_preprocess to be True.
6. If no parking policy is being applied, set implement_policy to be False.
7. Edit the setting EF_dir and base_lu to be the directory with the outputs from Estimates and Forecasts that the land use prep tool will process.
8. Edit the output_dir to the desired output location.
9. Update the name of config file with scenario year.
10. Open Anaconda prompt, navigate into the cloned repo and activate the environment and run run_landuse_preprocessing.bat. The files will be created in the specified output_dir

To easily run parking preprocessing

1. Create a directory to run in and clone the repo into the directory. Create folders in the clone called "parking_processed" and "parking_inputs"
2. Copy the the files auxiliary_columns.csv and micro_mobility_allyears.csv from T:\ABM\data\sr15_inputs\landuse_prep into the directory.
3. Copy the file 'mgra_parking_inventory.csv' from T:\ABM\data\sr15_inputs\landuse_prep\parking_inputs into the parking_inputs folder ("old" folder not needed).
4. Copy ParkingPolicies_[YEAR].csv from T:\ABM\data\sr15_inputs\landuse_prep\parking_outputs into the _processed folder.
5. Open up config_procpkg.yaml and set run_parking_inventory_preprocess and 'run_parking_space_estimation' to be True .
6. Edit the setting base_lu to be the directory with the outputs from Estimates and Forecasts that the land use prep tool will process.
7. Edit the setting bike_net and 'bike_node' to be the directory with the latest active transportation network.
8. Open Anaconda prompt, navigate into the cloned repo and create an Anaconda environment using the environment.yml file.
9. Activate the environment and run run_parking_preprocessing.bat. The files will be created in the specified parking_output_dir

To read PopulationSim outputs from Databricks

1. secrets.yml

Copy secrets.example.yml to secrets.yml and fill in the workspace URL, SQL warehouse id, catalog and schema. This file is included in the .gitignore to keep it private.

databricks:
  host: "<workspace-url>" # Databricks workspace URL
  warehouse_id: "<warehouse-id>" # SQL Warehouse connection details
  catalog: "<catalog>" # Catalog for PopulationSim output tables
  schema: "<schema>" # Schema for PopulationSim output tables

2. config_[YEAR].yaml

Set popsim_run_id to the PopulationSim run to read. popsim_landuse_table is the Databricks table the land use is read from (silver_mgra_based_input).

3. Sign in

When the land use is read from Databricks, a browser window opens to sign in with your Databricks account.