Scenario Space Exploration for Robust Energy Planning

This folder contains all the data related to the paper "Scenario Space Exploration for Robust Energy Planning".

Streamlit app

The interactive dashboard is deployed on Streamlit Community Cloud:

https://sserep.streamlit.app/

⚠️ Stability notice (RAM limits): The Streamlit Community Cloud instance has limited memory. Because the dashboard loads large model-output datasets into memory, the app may occasionally fail to start or restart unexpectedly (for example during initial data loading).

We’re actively working on a more stable deployment and improved, memory-efficient data loading. A more reliable version will be made available soon.

Contents

Parameter Space Files (Parameter_space/)

• parameter_space_LHS.xlsx - Parameter space definition used for Latin Hypercube Sampling (LHS)
• parameter_space_Morris.xlsx - Parameter space definition used for Morris sampling

Note: the Streamlit UI loads parameter-space definitions from UI/data/Original_data/Parameter space/.... The repo-root Parameter_space/ folder is provided as a convenient copy of the source spreadsheets.

Model Results (Model_results/)

LHS Results (model_results/LHS/)

Latin Hypercube Sampling (LHS) is used for the Delta sensitivity analysis method:

• GSA_Convergence_Analysis.csv - Convergence analysis for LHS sampling
• GSA_Delta.csv - Main Delta sensitivity analysis results
• GSA_Delta_*.csv - Delta analysis results for different sample sizes (400-4000)
• GSA_Delta_AllOutcomes_4000.csv - All outcomes Delta analysis (4000 samples)
• GSA_Delta_All_Re-Samples.csv - Re-sampling analysis results

Morris Results (model_results/Morris/)

Morris sampling is used for the Morris Elementary Effects method:

• GSA_Morris.csv - Main Morris sensitivity analysis results
• GSA_Morris_AllOutcomes.csv - Morris analysis for all outcomes

Energy System Model (IESA-Opt/)

The IESA-Opt energy system model used for scenario generation and analysis.

Requirements:

• AIMMS (free academic license available)
• Gurobi (free academic license available)

Documentation

• Workflow diagram.png - Visual representation of the scenario space exploration workflow
• README.md - This documentation file

Libraries and Tools Used

Sampling and Analysis

• SALib - Sensitivity Analysis Library for sampling methods and GSA calculations
• EMA Workbench - Exploratory Modeling and Analysis framework for scenario discovery

Dashboard

• Streamlit - Web application framework for the interactive dashboard

Methodology Notes

• LHS (Latin Hypercube Sampling): Used with the Delta sensitivity analysis method for variance-based sensitivity analysis

• Morris Sampling: Used with Morris Elementary Effects method for screening-level sensitivity analysis

• Feb 2026 update (UI): the dashboard is now Parquet-first for model outputs.

  • Required (default): UI/data/Generated_data/PPResults/<project>/<sample>/Model_Results_filtered.parquet
  • Optional: Model_Results.parquet (unfiltered)
  • Deprecated: “light/defaults” CSV outputs under UI/data/Generated_data/Defaults/...

• The sensitivity-analysis summary tables (e.g. GSA_*.csv) are still shipped/used as CSV.

For the exact expected directory structure and copy instructions, see UI/DATA_SETUP.md.

License

This project is licensed under the MIT License - see the LICENSE file for details.