Model Viewer App

Interactive visualisation of model predictions with optional uncertainty and validation overlays.
Supports 1D, 2D, and 3D views using Plotly.

Data expectations

• input_df (CSV) – rows are evaluation points; columns are continuous input variables (e.g., x,y,z).
• pred_df (CSV) – same number of rows as input_df; columns are output predictions (e.g., c, c2).
• unc_df (optional, CSV) – same rows/columns as pred_df; predictive 1σ.
• val_df (optional, CSV) – scattered validation points; must include the selected inputs and the selected output column.
• val_err_df (optional, CSV) – scattered validation errors; must include the selected inputs and an error column named <output>_err (percentage) or <output>.

Run locally

python -m venv .venv
source .venv/bin/activate     # (Windows: .venv\\Scripts\\activate)
pip install -r requirements.txt
streamlit run app.py