lets_plot.bistro.corr.corr_plot#

class lets_plot.bistro.corr.corr_plot(data, show_legend=True, flip=True, threshold=None)#

This class is intended to build correlation matrix plot.

Examples

1import numpy as np
2from lets_plot import *
3from lets_plot.bistro.corr import *
4LetsPlot.setup_html()
5np.random.seed(42)
6data = {var: np.random.poisson(size=10) for var in 'abcdef'}
7corr_plot(data).tiles().build()

1import numpy as np
2from lets_plot import *
3from lets_plot.bistro.corr import *
4LetsPlot.setup_html()
5np.random.seed(42)
6data = {var: np.random.uniform(size=10) for var in 'abcd'}
7corr_plot(data).tiles(type='upper', diag=True)\
8    .labels(type='upper', diag=True, map_size=True, color='black')\
9    .palette_RdBu().build()

1import numpy as np
2from lets_plot import *
3from lets_plot.bistro.corr import *
4LetsPlot.setup_html()
5np.random.seed(42)
6data = {var: np.random.normal(size=10) for var in 'abcdef'}
7corr_plot(data, flip=False, threshold=.4).points().labels()\
8    .palette_gradient(low='#d73027', mid='#ffffbf', high='#1a9850')\
9    .build()
__init__(data, show_legend=True, flip=True, threshold=None)#
Parameters:
datadict or Pandas or Polars DataFrame

Correlation matrix or data (correlation will be calculated for each variable pair). Data will be recognized as correlation matrix if it has a square shape and all values are in range -1.0..+1.0 or NaN.

show_legendbool, default=True

If True legend is shown.

flipbool, default=True

If True the y axis is flipped.

thresholdfloat, default=0.0

Minimal correlation abs value to be included in result. Accept values between 0 and 1.

points(type=None, diag=None)#

Method defines correlation matrix layer drawn by points to the plot.

Parameters:
type{‘upper’, ‘lower’, ‘full’}

Type of matrix. Default - contextual.

diagbool

Determines whether to fill the main diagonal with values or not. Default - contextual.

Returns:
corr_plot

Correlation plot specification.

labels(type=None, diag=None, map_size=None, color=None)#

Method defines correlation matrix layer drawn with geom_text to the plot.

Parameters:
type{‘upper’, ‘lower’, ‘full’}

Type of matrix. Default - contextual.

diagbool

Determines whether to fill the main diagonal with values or not. Default - contextual.

map_sizebool

If True, then absolute value of correlation is mapped to text size. If False - the text size is constant. Default - contextual.

colorstr

Set text color. For more info see https://lets-plot.org/python/pages/aesthetics.html#color-and-fill. Default - contextual.

Returns:
corr_plot

Correlation plot specification.

tiles(type=None, diag=None)#

Method defines correlation matrix layer drawn as square tiles to the plot.

Parameters:
type{‘upper’, ‘lower’, ‘full’}

Type of matrix. Default - contextual.

diagbool

Determines whether to fill the main diagonal with values or not. Default - contextual.

Returns:
corr_plot

Correlation plot specification.

palette_gradient(low, mid, high)#

Set scale_color_gradient2() and scale_fill_gradient() for corr plot.

Parameters:
lowstr

Color for low end of gradient (correlation -1).

midstr

Color for mid point (correlation 0).

highstr

Color for high end of gradient (correlation 1).

Returns:
corr_plot

Correlation plot specification.

palette_BrBG()#

Set scale_color_brewer() with BrBG palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

palette_PiYG()#

Set scale_color_brewer() with PiYG palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

palette_PRGn()#

Set scale_color_brewer() with PRGn palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

palette_PuOr()#

Set scale_color_brewer() with PuOr palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

palette_RdBu()#

Set scale_color_brewer() with RdBu palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

palette_RdGy()#

Set scale_color_brewer() with RdGy palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

palette_RdYlBu()#

Set scale_color_brewer() with RdYlBu palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

palette_RdYlGn()#

Set scale_color_brewer() with RdYlGn palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

palette_Spectral()#

Set scale_color_brewer() with Spectral palette for corr plot.

Returns:
corr_plot

Correlation plot specification.

build() PlotSpec#

This method creates PlotSpec object.

Returns:
PlotSpec

Plot specification.