Plotting Apps

These scripts live in apps/plot/ and provide quick command-line visualisations for common UQ data types.

plot_cov.py

Plot 2-D marginal covariance ellipses for a multivariate normal.

This script reads a mean vector and covariance matrix and produces pairwise 2-D covariance ellipse plots as well as a triangular grid of all pairs.

Example:

python plot_cov.py -m mean.txt -c cov.txt 0 1 2

Arguments:

Flag	Default	Description
positional	all	Indices of parameters to show.
`-m, --mean`	`mean.txt`	Mean file.
`-c, --cov`	`cov.txt`	Covariance file.

plot_ens.py

Plot an ensemble of output-data curves (spaghetti plot).

Outputs: ensemble.png

Example:

python plot_ens.py -y ytrain.dat

Arguments:

Flag	Default	Description
`-y, --ydata`	`ytrain.dat`	Output data file.

plot_pcoord.py

Plot parallel-coordinate diagrams for multivariate data.

Data are normalised to [-1, 1] before plotting. Optional label files allow colour-coding by group.

Outputs: pcoord_*.png

Example:

python plot_pcoord.py -x ptrain.txt -y ytrain.txt -e 5

Arguments:

Flag	Default	Description
`-x, --xdata`	`ptrain.txt`	Input data file.
`-y, --ydata`		Optional output data file.
`-o, --outnames_file`	`outnames.txt`	Output names file.
`-p, --pnames_file`	`pnames.txt`	Parameter names file.
`-e, --every`	1	Sample thinning factor.
`-l, --labels_file`		Label file for group colouring.
`-c, --ndcut`	0	Chunk size for splitting dimensions (0 = all).

plot_pdfs.py

Plot probability density functions from MCMC or other samples.

Supports triangular pair-plots, individual marginal PDFs (histograms or KDEs), burn-in trimming, thinning, prior-range overlays, and nominal-value markers.

Example:

python plot_pdfs.py -p pchain.dat -t tri -b 1000 -e 5

Arguments:

Flag	Default	Description
positional	all	Indices of parameters to show.
`-p, --samples_file`	`pchain.dat`	Samples file.
`-n, --names_file`		Parameter names file.
`-l, --nominal_file`		Nominal parameter values file.
`-g, --prange_file`		Prior range file.
`-t, --plot_type`	`tri`	`tri`, `ind`, or `inds`.
`-f, --pdf_type`	`hist`	`hist` or `kde`.
`-b, --burnin`	0	Burn-in samples to discard.
`-e, --every`	1	Thinning interval.

plot_xx.py

Plot pairwise scatter plots of input data, optionally colour-coded by label.

Outputs: xx_<dim1>_<dim2>.png

Example:

python plot_xx.py -x qtrain.txt -e 2

Arguments:

Flag	Default	Description
`-x, --xdata`	`qtrain.txt`	Input data file.
`-p, --pnames_file`	`pnames.txt`	Parameter names file.
`-e, --every`	1	Sample thinning factor.
`-l, --labels_file`		Label file for group colouring.

plot_yx.py

Plot outputs versus one input dimension at a time (linear and log scale).

Outputs: yx_<outname>.png, yx_<outname>_log.png

Example:

python plot_yx.py -x qtrain.txt -y ytrain.txt -c 4 -r 3

Arguments:

Flag	Default	Description
`-x, --xdata`	`qtrain.txt`	Input data file.
`-y, --ydata`	`ytrain.txt`	Output data file.
`-o, --outnames_file`	`outnames.txt`	Output names file.
`-p, --pnames_file`	`pnames.txt`	Parameter names file.
`-e, --every`	1	Sample thinning factor.
`-c, --cols`	4	Number of subplot columns.
`-r, --rows`	6	Number of subplot rows.

plot_yxx.py

Plot outputs versus pairs of inputs in a triangular layout, colour-coded by the output value. Useful for identifying interaction effects.

Outputs: yxx_<iout>.png

Example:

python plot_yxx.py -x qtrain.txt -y ytrain.txt -e 2

Arguments:

Flag	Default	Description
`-x, --xdata`	`qtrain.txt`	Input data file.
`-y, --ydata`	`ytrain.txt`	Output data file.
`-o, --outnames_file`	`outnames.txt`	Output names file.
`-p, --pnames_file`	`pnames.txt`	Parameter names file.
`-e, --every`	1	Sample thinning factor.