Introduction to Data Visualization with Python（datacamp）

• Customizing plots

  1. Plotting multiple graphs

     import matplotlib.pyplot as plt
     plt.plot(x, y1, 'red')
     plt.plot(x, y2, 'blue')
     plt.xlabel('x')
     plt.title('title')
     plt.show()
     
     # units between 0 and 1
     plt.axes([x_figure, y_figure, width, height])
     
     #subplot:nmk:n rows * m columns ,k subplot
     plt.subplot( n,m ,k)
     # Use plt.tight_layout() to improve the spacing between subplots
     plt.tight_layout()
     
     
     

  2. Customizing axes

     #Controlling axis extent
     plt.axis((xmin, xmax, ymin, ymax))
     
     plt.xlim((min, max))
     plt.ylin((min, max))
     
     #Equal scaling on x, y axes
     plt.axis('equal')
     

  3. Legends, annotations, and styles

     #Wrovide labels for overlaid points and curves
     #That could be :'upper left' or 'lower center' or 'best' or 'right'
     plt.legend(___)
     
     #Text labels and arrows using annotate() method
     plt.annotate(text_label, xy=(___),xytext=(___), arrowprops={'color':___})
     
     #Working with plot styles, such as'ggplot'
     plt.style.use(___)
     

• Plotting 2D arrays

  1. Working with 2D arrays

     #meshgrids:假设u是长度为m的向量，v是长度为n的向量，则最终生成的矩阵X和Y的维度都是 nm
     u = np.linspace(a, b, c)
     v = np.linespace(num1, num2, num3)
     X,Y = np.meshgrid(u, v)
     

  2. Visualizing bivariate functions

     plt.pcolor(___)
     plt.colorbar()
     plt.show()
     
     # gray color may
     plt.pcolor(___, cmap='gray')
     
     #Attention
     plt.pcolor(Z)
     plt.pcolor(X, Y, Z) # Axes determined by mesh grid arrays X,y
     
     
     #Contour plots
     plt.contour(___)
     plt.show()
     
     #Filled contour plots
     plt.contourf(X, Y, Z 30)
     plt.colorbar()
     plt.show()
     

  3. Visualizing bivariate distributions

     #Histograms in 2D
     #Common choices : rectangles & hexagons
     plt.hist2d(x, y, bins=(___,___))
     plt.colorbar()
     
     plt.hexbin(x, y, gridsize=(___,___))
     plt.colorbar()
     

  4. Working with images

     #Loading images
     img = plt.imread(___)
     plt.imshow(img)
     plt.axis('off')
     plt.show()
     
     #Reduction to gray-scale image
     collapsed = img.mean(axis=2)
     plt.set_cmap('gray')
     plt.imshow(collapsed, cmap='gray')
     plt.axis('off')
     plt.show()
     
     #Uneven samples
     uneven = collapsed[::4, ::2]
     plt.imshow(uneven)
     plt.axis('off')
     plt.show()
     #Adjusting aspect ration
     plt.imshow(uneven, aspect=2.0)
     plt.axis('off')
     plt.show()
     

• Statistical plots with Seaborn

  1. Visualizing Regressions

     #seaborn
     #hue, col, row: strings #定义数据子集的变量，并在不同的图像子集中绘制
     import pandas as pd
     import matplotlib.pyplot as plt
     import seaborn as sns
     tips = sns.load_dataset(___)
     sns.lmplot(x=___,y=___, data=___, hue=___, palette=___)
     plt.show()
     
     #residplot: similar arguments as lmplot() but more flexible
     sns.residplot((x=___,y=___, data=___)
           
     #sns:http://www.manongjc.com/article/52835.html
     #https://www.jianshu.com/p/c26bc5ccf604
     #https://blog.csdn.net/wuzlun/article/details/80324849
     

  2. Visualizing univariate distributions

     #Strip plot
     sns.stripplot(y=___, data=___)
     sns.stripplot(x=___,y=___, data=___)
     #Spreading out strip plot : size=___,jitter=True
     
     #Swarm plot
     sns.swarmplot(x=___,y=___, data=___)
     #Changing orientation : orient='h'
     
     #Violin plot
     sns.violinplot(x=___,y=___, data=___)Visualizing multivariate distributions
     

  3. Visualizing multivariate distributions

     #Joint plots
     sns.jointplot(x=___,y=___, data=___)
     #kde: kind='kde', 'scatter', 'reg', 'resid'
     
     #Pair plots
     sns.pairplot(___)
     
     #Heat maps
     sns,heatmap(___)
     

• Analyzing time series and images

  1. Visualizing times series

     #make formatted label
     labels = dates.striftime('%b %d')
     print (labels)
     ['Jan 01',...]
     
     #xtick
     plt.xticks(dates, labels, rotation=___)
     

  2. Time series with moving windows

     # A 'moving window' refers to a time interval in a time
     !!!!!!!!!!!!!!待补充
     

  3. Histogram equalization in images

     #Image histograms
     orig = plt.imread(___)
     pixels = orig.flatten() 	#flatten a 2D array into a 1D array
     plt.hist(pixels,___)
     
     #Rescaling the image
     minval, maxval = orig.min(), orig.max()
     rescaled = (225/(maxval-minval)*(pixels - minval))
     plt.imshow(rescaled)
     plt.axis('off')
     plt.show()
     
     #Image histogram & CDF
     plt.hist(pixels,___)
     plt.twinx()		#表示共享x轴,添加y轴的坐标轴
     plt.hist(pixels, cumulative=True)
     plt.show()
     
     #Equalizing intensity values
     new_pixels = np.interp(pixels, bins[:-1], orig_cdf*255)		#一维线性插值
     new = new_pixels.reshape(orig.shape)
     plt.imshow(new)
     plt.axis('off')
     plt.show()