Python Pandas MultiIndex Plot has incorrectly colored legend

Question

This seems to be a genereal and repeated (and unsolved) issue when trying to plot information from a Python Pandas MultiIndex.

The plotting code:

df.unstack(0).plot(x='SHIFT', y='INFORMATION_SURPLUS_PCT')

will render the following graph:

Unfortunately the color of the legend lines do not match the lines on the graph, I've also tried the following code to supply a legend but this still doesn't match up:

df.unstack(0).plot(x='SHIFT', y='INFORMATION_SURPLUS_PCT', legend=df.index.levels)

Example df:

       INFORMATION_SURPLUS_DIFF  INFORMATION_SURPLUS_PCT  \
   SYMBOL                                                         
   LBTYB  0                   0.000000                 0.000000   
          1                   0.015046                11.206891   
          2                   0.042249                31.468118   
          3                   0.005421                 4.037495   
          4                  -0.005569                 0.000000   
          5                  -0.069742                 0.000000   
          6                   0.036281                27.022972   
          7                   0.072626                54.093432   
          8                  -0.026966                 0.000000   
          9                  -0.048241                 0.000000   
          10                  0.145121               108.088780   
   MNST   0                   0.000000                 0.000000   
          1                   0.062962                 0.000000   
          2                   0.029280                 0.000000   
          3                   0.056574                63.569509   
          4                   0.088341                99.265331   
          5                   0.052083                 0.000000   
          6                   0.041253                46.353665   
          7                   0.058658                65.911282   
          8                   0.080748                90.732985   
          9                   0.031419                 0.000000   
          10                  0.055326                62.167628   

              MUTUAL_INFORMATION  SHIFT  
   SYMBOL                                
   LBTYB  0             0.134261      0  
          1             0.149307     -1  
          2             0.176510     -2  
          3             0.139681     -3  
          4             0.128692     -4  
          5             0.064518     -5  
          6             0.170542     -6  
          7             0.206887     -7  
          8             0.107294     -8  
          9             0.086020     -9  
          10            0.279381    -10  
   MNST   0             0.088995      0  
          1             0.151957     -1  
          2             0.118276     -2  
          3             0.145569     -3  
          4             0.177337     -4  
          5             0.141078     -5  
          6             0.130248     -6  
          7             0.147653     -7  
          8             0.169743     -8  
          9             0.120414     -9  
          10            0.144322    -10

Looks like multiple curves are getting plotted -- possibly on top of the originals. We'll need a minimal working example to help you. In particular, you'll need to show us some way to construct df, so that we can see what data is actually in there. — Mike
– Mike, Commented Mar 16, 2016 at 18:37

jezrael · Accepted Answer · 2016-03-16 19:29:55Z

2

I think you can try:

fig, ax = plt.subplots(figsize=(8,6))
for i, grp in df.groupby(level=0):
    grp.plot(x='SHIFT', y='INFORMATION_SURPLUS_PCT', label=str(i), ax=ax)

answered Mar 16, 2016 at 19:29

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

densekernel Over a year ago

This is definitely more elegant than the solution I've posted - Thanks once again!

densekernel · Accepted Answer · 2016-03-16 18:56:43Z

0

This code snippet will work, it involves grouping the MultiIndex and plotting to the same figure in a loop - Open for improvements:

grouped = df.groupby(level=0)
ax = plt.figure()
first = True
for i, group in grouped:
    print i
    if first:
        ax = group.plot(x='SHIFT', y='INFORMATION_SURPLUS_PCT', label=str(i))
        first = False
    else:
        group.plot(ax=ax, x='SHIFT', y='INFORMATION_SURPLUS_PCT', label=str(i))

New plot image:

answered Mar 16, 2016 at 18:56

densekernel

1,3613 gold badges14 silver badges21 bronze badges

Collectives™ on Stack Overflow

Python Pandas MultiIndex Plot has incorrectly colored legend

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related