Shulou(Shulou.com)06/01 Report--

This article is about how to use Matplotlib for data visualization. The editor thinks it is very practical, so share it with you as a reference and follow the editor to have a look.

01 Import Matplotlib

If you have installed the full Python Anaconda, then you have installed Matplotlib and are ready to start. Otherwise, you may want to visit the official website for installation instructions.

Http://matplotlib.org

Just as we use the acronym np for NumPy, we also use some standard abbreviations for Matplotlib imports:

Import matplotlib as mplimport matplotlib.pyplot as plt

Plt is one of our most commonly used interfaces.

02 generate a simple graph

Back to the point, let's create the first graphic.

Suppose we want to draw a simple diagram of the sinusoidal function sin (x). We want the function to find all the values on the x-axis (0 ≤ x ≤ 10). We will use NumPy's linspace function to create a linear space on the x-axis with x values from 0 to 10, a total of 100 sample points:

Import numpy as npx = np.linspace (0,10,100)

We can use the sin function of NumPy to find all the x values of the sin function and visualize the result by calling the plot function of plt:

Plt.plot (x, np.sin (x))

Have you tried it yourself? What happened? Did you find anything?

The problem is, depending on where you run the script, you may not see anything. Here are the possibilities that can be considered:

1. Draw from a .py script

If you are running matplotlib from a script, you only need to call plt, as follows:

Plt.show ()

After the call, the graph will be displayed!

two。 Draw from IPython shell

This is actually one of the most convenient ways to run matplotlib interactively. To display the drawing, you need to call the% matplotlib magic command after starting IPython:

% matplotlib Using matplotlib backend: Qt5Aggimport matplotlib.pyplot as plt

Then, all the diagrams are displayed automatically, so you don't have to call plt.show () every time.

3. Draw from Jupyter Notebook

If you view this code from a browser-based Jupyter Notebook, you need to use the same% matplotlib magic command. However, you can also choose to embed graphics directly into notebook, which has two possible outcomes:

% matplotlib notebook embeds the generated interaction diagram in notebook.

% matplotlib inline embeds the generated static graph in notebook.

We usually choose the inline option:

% matplotlib inline

Now, let's try again:

Plt.plot (x, np.sin (x))

The output given by the above command is shown in figure 2-4.

▲ figure 2-4 diagrams generated using the inline option

Later, if you want to save the chart, you can save it directly from the options in IPython or Jupyter Notebook:

Plt.savefig ('figures/02.03-sine.png')

Just make sure to use supported file suffixes, such as .jpg, .png, .tif, .svg, .eps, or .pdf.

After importing matplotlib, run plt.style.use (style_name), and you can change the style of the drawing. All available styles are listed in plt.style.available. For example, try plt.style.use ('fivethirtyeight'), plt.style.use (' ggplot'), or plt.style.use ('seaborn-dark'). For fun, you can run plt.xkcd () and then try to draw something else.

03 visualize the data of an external dataset

As the last test in this article, let's visualize some data from external datasets, such as scikit-learn 's digits dataset.

Specifically, we will need three visualization tools:

Scikit-learn for actual data

NumPy for data processing

Matplotlib

First, let's import all of these visualization tools:

Import numpy as np from sklearn import datasets import matplotlib.pyplot as plt matplotlib inline

The first step is to actually load the data:

Digits = datasets.load_digits ()

If we remember correctly, digits should have two different fields: one is the data field, which contains the actual image data, and the other is the target field, which contains the image label.

Instead of trusting our memories, let's look at digits objects. This is done by entering the field name, adding a period, and then pressing the Tab key-digits. To make it happen. This shows that the digits object also contains some other fields, such as a field named images. The two fields images and data seem to have different shapes:

Print (digits.data.shape) print (digits.images.shape)

Output result:

(1797, 64) (1797, 8, 8)

In both cases, the first dimension corresponds to the number of images in the dataset. But data arranges all the pixels in a large vector, while images retains the 8 × 8 spatial arrangement of each image.

Therefore, if we want to draw a single image, the images field may be more appropriate. First, grab an image from the dataset using NumPy's array cut:

Img = digits.images [0,:,:]

Here, we say that we want to grab the first row of the array with 1797 terms, as well as all the corresponding 8 × 8 pixels. Then we can use the imshow function of plt to draw the image:

Plt.imshow (img, cmap='gray') plt.savefig ('figures/02.04-digit0.png')

The output given by the above command is shown in figure 2-5. Notice that the image is blurred because we resized it to a larger size. The size of the original image is only 8 × 8.

▲ figure 2-5 example result of generating a single image

In addition, we can use the cmap parameter to specify a color map. By default, Matplotlib uses MATLAB's default color jet. However, for grayscale images, gray color images are more meaningful.

Finally, we can use the subplot function of plt to draw a set of digital samples. The subplot function, as in MATLAB, specifies the number of rows, columns, and the index of the current subgraph (starting at 1). We will use a for loop to traverse the first 10 images in the dataset, each with its own subgraph:

Plt.figure (figsize= (14,4)) for image_index in range (10): # images are 0-indexed, subplots are 1-indexed subplot_index = image_index + 1 plt.subplot (2,5, subplot_index) plt.imshow (digits.images [image _ index,:,:], cmap='gray')

The resulting output is shown in figure 2-6.

▲ figures 2-6 generate a set of subgraphs containing 10 numbers

Thank you for reading! This is the end of the article on "how to use Matplotlib for data visualization". I hope the above content can be of some help to you, so that you can learn more knowledge. if you think the article is good, you can share it for more people to see!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.