Drawing library Matplotlib styles and styles

1. Matplotlib style

In order to meet different application requirements ,Matplotlib  It contains  28  Different styles , When drawing , You can choose different drawing styles according to your needs . Use the following code to get  Matplotlib  All available styles in ：

import matplotlib as mplfrom matplotlib import pyplot as plt#  see  Matplotlib  Available drawing styles print(plt.style.available)

The available styles printed out are as follows ：

['Solarize_Light2', '_classic_test_patch', '_mpl-gallery', '_mpl-gallery-nogrid', 'bmh', 'classic', 'dark_background', 'fast', 'fivethirtyeight', 'ggplot', 'grayscale', 'seaborn', 'seaborn-bright', 'seaborn-colorblind', 'seaborn-dark', 'seaborn-dark-palette', 'seaborn-darkgrid', 'seaborn-deep', 'seaborn-muted', 'seaborn-notebook', 'seaborn-paper', 'seaborn-pastel', 'seaborn-poster', 'seaborn-talk', 'seaborn-ticks', 'seaborn-white', 'seaborn-whitegrid', 'tableau-colorblind10']

Use  plt.style.use("style")  You can change the default drawing style , among  style  by  Matplolib  Styles available in ：

import matplotlib as mplfrom matplotlib import pyplot as pltimport numpy as np#  Set the default drawing style to  seaborn-darkgridplt.style.use("seaborn-darkgrid")  x = np.linspace(0, 2 * np.pi, 100)y = np.cos(x)plt.plot(x, y)plt.show()

You can see , And in 《Matplotlib Graph drawing 》 Compared with the same figure drawn in , Its style is quite different . besides ,Maplotlib  There are also some very interesting styles in , For example, hand-painted style ：

import matplotlib as mplfrom matplotlib import pyplot as pltimport numpy as np#  Enable freehand style plt.xkcd()x = np.linspace(0, 2 * np.pi, 100)y = np.cos(x)plt.plot(x, y)plt.show()

You can see , The method of using hand drawn style is very simple , Just before drawing , call  plt.xkcd()  Function ：

import matplotlib as mplfrom matplotlib import pyplot as pltimport numpy as np#  Enable freehand style plt.xkcd()data = [[10., 20., 30., 20.],[40., 25., 53., 18.],[6., 22., 52., 19.]]x = np.arange(4)plt.bar(x + 0.00, data[0], color = 'b', width = 0.25)plt.bar(x + 0.25, data[1], color = 'g', width = 0.25)plt.bar(x + 0.50, data[2], color = 'r', width = 0.25)plt.show()

except  Matplotlib  In addition to its own preset style , We can also customize colors and styles to build custom styles that meet our needs .

2. Use custom colors

2.1 Custom color

Matplotlib  There are many ways to define colors in , Common methods include ：

1. A triple (Triplets)： Color can be described as a real triplet , That is, the color of red 、 blue 、 Green weight , Each of these components is in  [0,1]  Within the interval . therefore ,(1.0, 0.0, 0.0)  It means pure red , and  (1.0, 0.0, 1.0)  It means pink .

2. Four tuple (Quadruplets)： Their first three elements have the same definition as triples , The fourth element defines the transparency value . This value is also  [0,1]  Within the interval . When rendering graphics to a picture file , Use transparent colors to mix the drawing with the background .

3. Predefined name ：Matplotlib  Standard  HTML  The color name is interpreted as the actual color . for example , character string  red  It can be expressed as red . At the same time, some colors have concise aliases , As shown in the following table ：

   Alias	Color	Show
   b	blue
   g	green
   r	red
   c	cyan
   m	magenta
   y	yellow
   k	black
   w	white

4. HTML  Color string ：Matplotlib  Can be  HTML  The color string is interpreted as the actual color . These strings are defined as  #RRGGBB, among  RR、GG and  BB  They are red with hexadecimal encoding 、 Green and blue components .

5. Grayscale string ：Matplotlib  Interprets the string representation of floating-point values as grayscale , for example  0.75  Indicates medium light gray .

2.2 Draw a graph with custom colors

By setting  plt.plot()  The parameters of the function  color ( Or equivalent abbreviation is  c), You can set the color of the curve , As shown below ：

import numpy as npimport matplotlib.pyplot as pltdef pdf(x, mu, sigma):    a = 1. / (sigma * np.sqrt(2. * np.pi))    b = -1. / (2. * sigma ** 2)    return a * np.exp(b * (x - mu) ** 2)x = np.linspace(-6, 6, 1000)for i in range(5):    samples = np.random.standard_normal(50)    mu, sigma = np.mean(samples), np.std(samples)    plt.plot(x, pdf(x, mu, sigma), color = str(.15*(i+1)))plt.plot(x, pdf(x, 0., 1.), color = 'k')plt.plot(x, pdf(x, 0.2, 1.), color = '#00ff00')plt.plot(x, pdf(x, 0.4, 1.), color = (0.9,0.9,0.0))plt.plot(x, pdf(x, 0.4, 1.), color = (0.9,0.9,0.0,0.8))plt.show()

2.3 Draw a scatter chart with custom colors

You can control the color of the scatter chart in the same way as you control the graph , There are two forms available ：

1. Use the same color for all points ： All points will be displayed in the same color .
2. Define different colors for each point ： Provide different colors for each point .

2.3.1 Use the same color for all points

Using two sets of points extracted from binary Gaussian distribution  y_1  and  y_2, The color of the midpoint in each group is the same ：

import numpy as npimport matplotlib.pyplot as plty_1 = np.random.standard_normal((150, 2))y_1 += np.array((-1, -1)) # Center the distrib. at <-1, -1>y_2 = np.random.standard_normal((150, 2))y_2 += np.array((1, 1)) # Center the distrib. at <1, 1>plt.scatter(y_1[:,0], y_1[:,1], color = 'c')plt.scatter(y_2[:,0], y_2[:,1], color = 'b')plt.show()

2.3.2 Define different colors for each point

We sometimes encounter such drawing scenes , You need to draw different colors for different categories of points , To observe the differences between different categories . With Fisher’s iris Data sets For example , The data in its dataset is similar to the following ：

5.0,3.3,1.4,0.2,Iris-setosa7.0,3.2,4.7,1.4,Iris-versicolo

Each point of the dataset is stored in a comma separated list . The last column gives the label of each point ( Tags contain three categories ：Iris-virginica、Iris-versicolor  and  Iris-Vertosa). In the example , The color of these points will depend on their labels , As shown below ：

import numpy as npimport matplotlib.pyplot as pltlabel_set = (    b'Iris-setosa',    b'Iris-versicolor',    b'Iris-virginica',)def read_label(label):    return label_set.index(label)data = np.loadtxt('iris.data', delimiter = ',', converters = { 4 : read_label })color_set = ('c', 'y', 'm')color_list = [color_set[int(label)] for label in data[:,4]]plt.scatter(data[:,0], data[:,1], color = color_list)plt.show()

For three possible labels , We specify a unique color , Color in  color_set  In the definition of , Tag in  label_set  In the definition of .label_set  No  i  A label with  color_set  No  i  Associated with two colors . Then we use them to convert the label list into a color list  color_list. Then just call  plt.scatter()  All points and their colors can be displayed at once .
In addition to the above methods , We can also call... Separately by calling three different categories  plt.scatter()  To achieve , But this will require more code . Another thing to note ： If two points may have the same coordinates , But there are different labels , The color displayed will be the color of the point drawn after , We can also use color values with transparent channels , Used to display overlapping points .

2.4 Use custom colors for the edges of data points in a scatter chart

And  color  The color of the parameter control point is the same , have access to  edgecolor  Parameter controls the color of the edge of the data point , You can set the same color for the edges of each point ：

import numpy as npimport matplotlib.pyplot as pltdata = np.random.standard_normal((100, 2))plt.scatter(data[:,0], data[:,1], color = '1.0', edgecolor='r')plt.show()

In addition to setting the same color for the edges of each point , You can also set the color of the edges of each point as described in the section defining different colors for each point .

2.5 Draw a bar chart with custom colors

Controls the colors used to draw bar charts, which work the same way as curves and scatter charts , That is, through optional parameters  color：

import numpy as npimport matplotlib.pyplot as pltw_pop = np.array([5., 30., 45., 22.])m_pop = np.array( [5., 25., 50., 20.])x = np.arange(4)plt.barh(x, w_pop, color='m')plt.barh(x, -m_pop, color='c')plt.show()

Again , Use  plt.bar()  and  plt.barh()  The function customizes the color of the bar graph to work in the same way as  plt.scatter()  Exactly the same , Just set the optional parameters  color, At the same time, you can also use parameters  edgecolor  Controls the color of the bar edges .`

import numpy as npimport matplotlib.pyplot as pltvalues = np.random.random_integers(99, size = 50)color_set = ('c', 'm', 'y', 'b')color_list = [color_set[(len(color_set) * val) // 100] for val in values]plt.bar(np.arange(len(values)), values, color = color_list)plt.show()

2.6 Draw pie charts with custom colors

The method of customizing the color of pie chart is similar to that of bar chart ：

import numpy as npimport matplotlib.pyplot as pltcolor_set = ('c', 'm', 'y', 'b')values = np.random.rand(6)plt.pie(values, colors = color_set)plt.show()

Pie charts are acceptable  colors  Parameters ( Be careful , Here is  colors, Not in  plt.plot()  Used in  color ) A list of colors . however , If the number of colors is less than the number of elements in the input value list , that  plt.pie()  The colors in the color list will be recycled . In the example , Use a list of four colors , To color a pie chart with six values , Two of these colors will be used twice .

2.7 Draw a box diagram with custom colors

Use the following code , Modify the line color in the box drawing ：

import numpy as npimport matplotlib.pyplot as pltvalues = np.random.randn(100)b = plt.boxplot(values)for name, line_list in b.items():    for line in line_list:        line.set_color('m')plt.show()

2.8 Use color mapping to draw a scatter plot

If you want to use multiple colors in the drawing , Defining each color one by one is not the best solution , Color mapping can solve this problem . Color mapping uses a variable to correspond to a value ( Color ) The continuous function of defines the color .Matplotlib  Several common color mappings are provided ; Most are continuous color gradients . Color mapping in  matplotib.cm  Defined in module , Provides functions for creating and using color maps , It also provides predefined color mapping options .
function  plt.scatter()  Accept  color  List of values of parameters , When provided  cmap  Parameter values , These values will be interpreted as the index of the color map ：

import numpy as npimport matplotlib.cm as cmimport matplotlib.pyplot as pltn = 256angle = np.linspace(0, 8 * 2 * np.pi, n)radius = np.linspace(.5, 1., n)x = radius * np.cos(angle)y = radius * np.sin(angle)plt.scatter(x, y, c = angle, cmap = cm.hsv)plt.show()

stay  matplotlib.cm  A large number of predefined color mappings are provided in the module , among  cm.hsv  Contains a full spectrum of colors .

2.9 Use color mapping to draw bar charts

plt.scatter()  Function has built-in support for color mapping , Other drawing functions also have built-in support of color mapping . however , Some functions ( Such as  plt.bar() ) No built-in support for color mapping , under these circumstances ,Matplotlib  You can explicitly generate colors from color maps ：

import numpy as npimport matplotlib.cm as cmimport matplotlib.pyplot as pltimport matplotlib.colors as colvalues = np.random.random_integers(99, size = 50)cmap = cm.ScalarMappable(col.Normalize(0, 99), cm.binary)plt.bar(np.arange(len(values)), values, color = cmap.to_rgba(values))plt.show()

The above code first creates a color map  cmap, So that  [0, 99]  Values in the range are mapped to  matplotlib.cm.binary  The color of the . then , function  cmap.to_rgba  Convert value list to color list . therefore , Even though  plt.bar()  There is no built-in color mapping support , But you can still use uncomplicated code to realize color mapping .

2.10 Create a custom color scheme

Matplotlib  The default color used is mainly considered for printed documents or publications . therefore , By default , The background is white , And the label 、 Axes and other annotations are displayed in black , In some different use environments , The color scheme we may need to use ; for example , Set the graphic background to black , The annotation is set to white .
stay  Matplotlib  in , Various objects ( Such as axis 、 Graphics and labels ) Can be modified separately , But changing the color configuration of these objects one by one is not the best solution . We can use it in 《Matplotlib Installation and configuration 》 Methods introduced , Change all objects by modifying the configuration set , To configure its default color or style ：

import numpy as npimport matplotlib as mplfrom matplotlib import pyplot as pltmpl.rc('lines', linewidth = 2.)mpl.rc('axes', facecolor = 'k', edgecolor = 'w')mpl.rc('xtick', color = 'w')mpl.rc('ytick', color = 'w')mpl.rc('text', color = 'w')mpl.rc('figure', facecolor = 'k', edgecolor ='w')mpl.rc('axes', prop_cycle = mpl.cycler(color=[(0.1, .5, .75),(0.5, .5, .75)]))x = np.linspace(0, 7, 1024)plt.plot(x, np.sin(x))plt.plot(x, np.cos(x))plt.show()

3. Use custom styles

3.1 Control line styles and lineweights

In practice , Except for color , In most cases, we also need to control the line style of graphics , Add variety to line styles .

3.1.1 Line style

import numpy as npimport matplotlib.pyplot as pltdef gaussian(x, mu, sigma):    a = 1. / (sigma * np.sqrt(2. * np.pi))    b = -1. / (2. * sigma ** 2)    return a * np.exp(b * (x - mu) ** 2)x = np.linspace(-6, 6, 1024)plt.plot(x, gaussian(x, 0., 1.), color = 'y', linestyle = 'solid')plt.plot(x, gaussian(x, 0., .5), color = 'c', linestyle = 'dashed')plt.plot(x, gaussian(x, 0., .25), color = 'm', linestyle = 'dashdot')plt.show()

Use  plt.plot()  Of  linestyle  Parameters to control the style of the curve , Other available line styles include ：solid、dashed、dotted、dashdot. Again , Line style settings are not limited to  plt.plot(), This parameter can be used for any drawing composed of lines , It can also be said that  linestyle  Parameters are available for all commands involving line rendering . for example , You can modify the line style of the bar chart ：

import numpy as npimport matplotlib.pyplot as pltn = 10a = np.random.random(n)b = np.random.random(n)x = np.arange(n)plt.bar(x, a, color='c')plt.bar(x, a+b, bottom=a, color='w', edgecolor='black', linestyle = 'dashed')plt.show()

Because in the bar chart 、 Pie chart and other graphics , The default border color is white , So to show these lines on a white background , Need to pass through  edgecolor  Parameter to change the default color of the edge line .

3.1.2 Line width

Use  linewidth  Parameter can change the thickness of the line . By default ,linewidth  Set to 1 A unit of . Using the thickness of lines can visually emphasize a specific curve .

import numpy as npimport matplotlib.pyplot as pltdef gaussian(x, mu, sigma):    a = 1. / (sigma * np.sqrt(2. * np.pi))    b = -1. / (2. * sigma ** 2)    return a * np.exp(b * (x - mu) ** 2)x = np.linspace(-6, 6, 1024)for i in range(64):    samples = np.random.standard_normal(50)    mu, sigma = np.mean(samples), np.std(samples)    plt.plot(x, gaussian(x, mu, sigma), color = '.75', linewidth = .5)plt.plot(x, gaussian(x, 0., 1.), color = 'c', linewidth = 3.)plt.show()

3.2 Control fill patterns

Matplotlib  A hatch pattern is provided to fill the plane . These patterns , It plays an important role in graphics containing only black and white .

import numpy as npimport matplotlib.pyplot as pltn = 10a = np.random.random(n)b = np.random.random(n)x = np.arange(n)plt.bar(x, a, color='w', hatch='x', edgecolor='black')plt.bar(x, a+b, bottom=a, color='w', edgecolor='black', hatch='/')plt.show()

Functions with filling and rendering ( Such as  plt.bar() ) Optional parameters can be used  hatch  Control fill patterns , Optional values for this parameter include ：/, \, |, -, +, x, o, O,.  and  *, Each value corresponds to a different hatch pattern ;edgecolor  Parameters can be used to control the color of the hatch pattern .

3.3 Control marks

3.3.1 Controls the marker style

stay 《Matplotlib Graph drawing 》 in , We've learned how to draw curves , And understand that the curve is composed of connecting lines between points ; Besides , The scatter chart shows each point in the dataset . and  Matplotlib  Offers a variety of shapes , You can replace the style of points with other types of markers .
There are several ways to specify tags ：

1. Predefined Tags ： Predefined shapes , Expressed as  [0, 8]  An integer within a range or some predefined string .
2. Vertex list ： List of value pairs , Coordinates used as the shape path .
3. Regular polygon ： Express  N  Triples of regular polygons with edges  (N, 0, angle) , among  angle  For the angle of rotation .
4. Star polygon ： It is represented as a triple  (N, 1, angle), representative  N  Side positive star , among  angle  For the angle of rotation .

import numpy as npimport matplotlib.pyplot as plta = np.random.standard_normal((100, 2))a += np.array((-1, -1))b = np.random.standard_normal((100, 2))b += np.array((1, 1))plt.scatter(a[:,0], a[:,1], color = 'm', marker = 'x')plt.scatter(b[:,0], b[:,1], color = 'c', marker = '^')plt.show()

You can see , Use  marker  Parameters , You can specify different tags for each data set .
We have learned how to define different colors for each point in the scatter diagram , What if we need to define different styles for each point ？ The problem lies in , And  color  Different parameters ,marker  Input styles are not accepted as parameter lists . therefore , We can't achieve  plt.scatter()  To display multiple point sets with different markers . The solution is , Separate the data points of each type into different sets , And call... Separately for each collection  plt.scatter()  call ：

import numpy as npimport matplotlib.pyplot as pltlabel_list = (    b'Iris-setosa',    b'Iris-versicolor',    b'Iris-virginica',)colors = ['c','y','m']def read_label(label):    return label_list.index(label)data = np.loadtxt('iris.data', delimiter = ',', converters = { 4 : read_label })marker_set = ('^', 'x', '.')for i, marker in enumerate(marker_set):    data_subset = np.asarray([x for x in data if x[4] == i])    plt.scatter(data_subset[:,0], data_subset[:,1], color = colors[i], marker = marker)plt.show()

about  plt.plot(), You can also access tag styles using the same tag parameters . When data points are dense , Using markers to display each point will cause confusion in the picture , therefore  Matplotlib  Provides  markevery  Parameters , Allow every  N  A dot displays a marker ：

import numpy as npimport matplotlib.pyplot as pltx = np.linspace(-6, 6, 1024)y_1 = np.sinc(x)y_2 = np.sinc(x) + 1plt.plot(x, y_1, marker = 'x', color = '.75')plt.plot(x, y_2, marker = 'o', color = 'k', markevery = 64)plt.show()

3.3.1 Control tag size

The size of the tag is optional  s  Control ：

import numpy as npimport matplotlib.pyplot as plta = np.random.standard_normal((100, 2))a += np.array((-1, -1))b = np.random.standard_normal((100, 2))b += np.array((1, 1))plt.scatter(a[:,0], a[:,1], c = 'm', s = 100.)plt.scatter(b[:,0], b[:,1], c = 'c', s = 25.)plt.show()

The size of the tag is determined by  plt.scatter()  Parameters of  s  Set up , However, it should be noted that it sets the surface area ratio of the mark rather than the radius .plt.scatter()  The function can also accept a list as  s  Parameter input , It represents the size of each point ：

import numpy as npimport matplotlib.pyplot as pltm = np.random.standard_normal((1000, 2))r_list = np.sum(m ** 2, axis = 1)plt.scatter(m[:, 0], m[:, 1], c = 'w', edgecolor='c', marker = 'o', s = 32. * r_list)plt.show()

plt.plot()  Functions are allowed in  markersize ( Or abbreviated as  ms ) Parameter to change the size of the tag , However, this parameter does not accept a list as input .

3.3.3 Create custom tags

although  Matplotlib  A variety of marker shapes are available . But in some cases, we may still not find a shape suitable for specific needs . for example , We may want to use company logos, etc. as shapes .
stay  Matplotlib  in , Describe the shape as a path —— The connection of a series of points . therefore , If we want to define our own tag shape , A series of points must be provided ：

import numpy as npimport matplotlib.path as mpathfrom matplotlib import pyplot as pltshape_description = [    ( 1., 2., mpath.Path.MOVETO),    ( 1., 1., mpath.Path.LINETO),    ( 2., 1., mpath.Path.LINETO),    ( 2., -1., mpath.Path.LINETO),    ( 1., -1., mpath.Path.LINETO),    ( 1., -2., mpath.Path.LINETO),    (-1., -2., mpath.Path.LINETO),    (-1., -1., mpath.Path.LINETO),    (-2., -1., mpath.Path.LINETO),    (-2., 1., mpath.Path.LINETO),    (-1., 1., mpath.Path.LINETO),    (-1., 2., mpath.Path.LINETO),    ( 0., 0., mpath.Path.CLOSEPOLY),]u, v, codes = zip(*shape_description)my_marker = mpath.Path(np.asarray((u, v)).T, codes)data = np.random.rand(8, 8)plt.scatter(data[:,0], data[:, 1], c = 'm', marker = my_marker, s = 75)plt.show()

Through the above example , You can see , All marked  plt  Drawing functions have an optional parameter  marker, The parameter value can be predefined  Matplotlib  Mark , It can also be a custom path instance , The path object is in  matplotlib.path  Defined in module .
Path  Object's constructor takes a list of coordinates and a list of instructions as input ; One command per coordinate , Use a list to fuse coordinates and instructions , Then pass the coordinate list and instructions to the path constructor , As shown below ：

u, v, codes = zip(*shape_description)my_marker = mpath.Path(np.asarray((u, v)).T, codes)

The shape is described by the movement of the cursor ：

• MOVETO： This command moves the cursor to the specified coordinates , Don't draw lines .
• LINETO： This draws a line between the current point of the cursor and the target point , Move the cursor to the target point .
• CLOSEPOLY： This command is only used to close the path , Each shape ends with this indication .

Theoretically , Any shape is possible , We just need to describe its path . But in practice , If you want to use complex shapes , It is better to carry out the conversion work in advance .

Related links

Matplotlib Installation and configuration
Matplotlib Quick start
Matplotlib Graph drawing