Cornerplots

Simple library to create corner plots in matplotlib. Corner plots provide a simple way to visualize multidimensional data, with each dimension plotted against every other dimension. The main goal is to let the user specify data and/or model predictions in arbitrary subsets of the multi-dimensional space and let the code take care of where to plot these values. Originally inspired by corner.py.

Simple example

You might have one set of (w, x, y) data, another set of (x, y, z) data, and a model predicting (w, x, y, z) trajectories.

import numpy as np

t = np.linspace(0, 1)
model = dict(
    w = 3 * t + 1,
    x = 5 * t,
    y = 7 * t**.5,
    z = 3 * t**1.5,
    )

red_data = dict(
    w = np.array([1, 2, 3, 4]),
    x = np.array([0, 2, 4, 5]),
    y = np.array([1, 4, 6, 8]),
    )

blue_data = dict(
    x = np.array([1, 2, 5]),
    y = np.array([3, 5, 7]),
    z = np.array([0, 1, 3]),
    )

In that case, you could plot everything by simply calling:

from cornerplots import Cornerplots
import matplotlib.pyplot as ppl

fig = ppl.figure(figsize = (5,5))

cp = Cornerplots(
    fields = ['w', 'x', 'y', 'z'],
    labels = ['W-value', 'X-value', 'Y-value', 'Z-value'],
    grid_kwargs = {'alpha': 0.25, 'lw': 0.7},
    )

cp.plot(model,     'k-', lw = 1, dashes = (6,2,2,2), label = 'model')
cp.plot(red_data,  'ws', mec = 'r', mew = 1, ms = 5, label = 'red data')
cp.plot(blue_data, 'wD', mec = 'b', mew = 1, ms = 5, label = 'blue data')

cp.legend()

ppl.savefig('example.png')