This repo contains the volume forecasts for externals for the Wasatch Front Travel Demand
Model. The pipeline is a single Quarto document, index.qmd, rendered to docs/
(published at https://wfrcanalytics.github.io/TDM-INP-External-Forecasts/).
index.qmd is a single-document port of five sequential notebooks that used to be the
pipeline. Those notebooks are kept in the repo root as archived/historical reference
only — they are excluded from Quarto's render (see _quarto.yml) and are no longer run
or maintained:
1-Get-Historic-AADT.ipynb: The Average Annual Daily Traffic (AADT) historic data from the Utah Department of Transportation is matched to each external.
2-Prepare-Previous-Forecasts.ipynb: Forecasts from previous model versions are gathered and processed to inform forecasting.
3-Prepare-Linear-Forecasts.ipynb: Linear forecasts are created off an extrapolated linear least-square regression fit of the historic AADT. These trend lines are created for a user-defined set of year ranges, eg. 2011-2023 linear forecast.
4-Finalize-Forecasts.ipynb: Using a series of charts, forecaster determines the linear forecasts to use for each external, and then defines further manual adjustments as needed.
5-Export-Data.ipynb: The inputs for the externals for the TDM are generated.
This project uses uv to manage the Python environment (Python 3.11, pinned in .python-version).
uv sync
This creates a .venv/ and installs everything from pyproject.toml/uv.lock. To run a notebook's dependencies, prefix commands with uv run, e.g. uv run jupyter lab. A Jupyter kernel for this environment can be registered with:
uv run python -m ipykernel install --user --name tdm-inp-external-forecasts --display-name "TDM-INP-External-Forecasts (uv)"
externals-app/ is deployed separately to Posit Connect and keeps its own requirements.txt; it is not part of this uv project.