Skip to content

Use cases#

GloBI produces detailed building energy outputs, for example EnergyAndPeak.pq. You can use these results for many types of analysis: policy and retrofit comparisons, adoption and willingness-to-pay (WTP) models, emissions accounting, and other fleet-scale workflows.

Why this is a separate repository#

The WTP workflow needs many inputs that GloBI does not make: retrofit cost and incentive tables, energy prices, adoption curves, emissions-factor trajectories, demographic data for propensity models, and Monte Carlo settings. A separate repository, globi-use-cases, holds these tools. This keeps the GloBI install small, and lets the analyses change on their own release cycle.

Other downstream use cases are in the same repository, with the analysis tools, sample inputs in data/inputs/, and the Streamlit apps.

What you will find there#

The globi-use-cases repository has tools and a Streamlit app for:

  1. Energy and policy impacts: compare baseline and retrofit scenario files (one pair for each year) for energy and peak differences.
  2. Propensity: calculate the acceptance probability of each building (residential logit with census demographics; commercial NPV threshold).
  3. Uptake: apply adoption curves from JSON, so that buildings adopt over time from propensity and curve targets.
  4. Emissions: combine the adopted floor area and fuel mix with emissions-factor trajectories that you can edit.

In the main app, you upload results, set the configuration, select adoption curves, edit emissions, and run the full pipeline with charts and tables. The sample inputs in data/inputs/ let you use the app without new GloBI runs.

For the full feature list, the scenario editor, Docker options, the data layout and the split_scenarios.py tool, see the README of that repository.

Download the repository#

Clone the repository with Git. Git lets you get updates later with git pull.

git clone https://github.com/daryaguettler/globi-use-cases.git
cd globi-use-cases

You can also download a ZIP file from the Code menu on GitHub. Extract it, then cd into the extracted folder.

Requirements#

  • Python 3.11+
  • uv (package manager)
  • Docker Desktop, to run the app in Docker with make run

Set up the environment#

To run the app on your computer (not in Docker), install the Python environment from the repository root:

uv sync

The Makefile sets PYTHONPATH=src, so that the app and use_cases imports work. Use the make targets below, or set PYTHONPATH yourself.

Run the analysis app#

To run the app in Docker and open it in your browser:

make run

To run the app on your computer without Docker:

make run-local

make run-local is the same as:

PYTHONPATH=src uv run streamlit run src/app/wtp_app.py

Open the URL that Streamlit prints. The default is http://localhost:8502, so the app can run at the same time as the GloBI results dashboard (port 8501).

The workflow in the app:

  1. Upload pairs of baseline and scenario EnergyAndPeak.pq files.
  2. Set the costs and options in the sidebar.
  3. Select the adoption curves and emissions trajectories.
  4. Run the pipeline and look at the results.

To edit adoption curves and emissions trajectories in a separate app, run make editor. To build and run the Docker image directly, run make docker-build, then make docker-run.