Legend table layout for matplotlib

Question

I have a 9-line graph representing datasets with two variable parameters, say f_11, f_12, f_13, ..., f_33. To make the graph (bit) clearer, I encode the first parameter as the color of the line, and the second as linestyle (so f_11 is red and dashed, f12 is red and dashed, f21 is green and dashed, f22 is green and dashed, etc.) . So for the legend, I would like to make a 3x3 table looking like

| value1 | value2 | value3 --------------------------------- value1 | value2 | <artists go there> value3 | 

Is there a way to do this with matplotlib? The idea would be to make this box with LaTeX, but I need a way to capture the legendary artists in the right position.

Thanks!

(crossover from matplotlib users)

Answer 1

Not a very simple question, but I figured it out. The trick I'm using is to initialize an empty rectangle that acts like a handle. These additional empty descriptors are used to build the table. I get rid of extra space using handletextpad :

 import numpy import pylab import matplotlib.pyplot as plt from matplotlib.patches import Rectangle fig = plt.figure() ax = fig.add_subplot(111) im1 ,= ax.plot(range(10), pylab.randn(10), "r--") im2 ,= ax.plot(range(10), pylab.randn(10), "g--") im3 ,= ax.plot(range(10), pylab.randn(10), "b--") im4 ,= ax.plot(range(10), pylab.randn(10), "r.") im5 ,= ax.plot(range(10), pylab.randn(10), "g.") im6 ,= ax.plot(range(10), pylab.randn(10), "b.") im7 ,= ax.plot(range(10), pylab.randn(10), "r^") im8 ,= ax.plot(range(10), pylab.randn(10), "g^") im9 ,= ax.plot(range(10), pylab.randn(10), "b^") # create blank rectangle extra = Rectangle((0, 0), 1, 1, fc="w", fill=False, edgecolor='none', linewidth=0) #Create organized list containing all handles for table. Extra represent empty space legend_handle = [extra, extra, extra, extra, extra, im1, im2, im3, extra, im4, im5, im6, extra, im7, im8, im9] #Define the labels label_row_1 = [r"$f_{i,j}$", r"$i = 1$", r"$i = 2$", r"$i = 3$"] label_j_1 = [r"$j = 1$"] label_j_2 = [r"$j = 2$"] label_j_3 = [r"$j = 3$"] label_empty = [""] #organize labels for table construction legend_labels = numpy.concatenate([label_row_1, label_j_1, label_empty * 3, label_j_2, label_empty * 3, label_j_3, label_empty * 3]) #Create legend ax.legend(legend_handle, legend_labels, loc = 9, ncol = 4, shadow = True, handletextpad = -2) plt.show() 

Answer 2

If you need to create a readable table, you can use PrettyTable module

For example in python2.7 this code

 from prettytable import PrettyTable x = PrettyTable(["City name", "Area", "Population", "Annual Rainfall"]) x.align["City name"] = "l" # Left align city names x.padding_width = 1 # One space between column edges and contents (default) x.add_row(["Adelaide",1295, 1158259, 600.5]) x.add_row(["Brisbane",5905, 1857594, 1146.4]) x.add_row(["Darwin", 112, 120900, 1714.7]) x.add_row(["Hobart", 1357, 205556, 619.5]) x.add_row(["Sydney", 2058, 4336374, 1214.8]) x.add_row(["Melbourne", 1566, 3806092, 646.9]) x.add_row(["Perth", 5386, 1554769, 869.4]) print x 

May produce this conclusion:

 +-----------+------+------------+-----------------+ | City name | Area | Population | Annual Rainfall | +-----------+------+------------+-----------------+ | Adelaide | 1295 | 1158259 | 600.5 | | Brisbane | 5905 | 1857594 | 1146.4 | | Darwin | 112 | 120900 | 1714.7 | | Hobart | 1357 | 205556 | 619.5 | | Sydney | 2058 | 4336374 | 1214.8 | | Melbourne | 1566 | 3806092 | 646.9 | | Perth | 5386 | 1554769 | 869.4 | +-----------+------+------------+-----------------+