Scatter plot legend shows only one variable with color -Pandas Seaborn

Question

I am trying to plot a scatter plot in pandas with seaborn package. I want both of my variables to show in legend, but I am only getting one. Following is a what I did:

import pandas as pd
import numpy as np
from matplotlib import pyplot as plt
import seaborn as sns 

df = pd.DataFrame(np.random.randn(100, 6), columns=['a', 'b', 'c', 'd', 'e', 'f'])

I plotted like this,

plt.scatter(df['a'],df['b'], color = ['red', 'blue'])
plt.legend(loc = 'best')
plt.show()

I am getting image like this,

As you can see, I do not see a column/object with blue color. What mistake am I making here? I searched on this, no luck yet. Any suggestion would be appreciated.

you are plotting points with `(x, y)` coordinates where `x=df['a']` and `y=df['b']` - is that what you want? — MaxU - stand with Ukraine, Aug 18 '17 at 21:04
@MaxU yes, that is what I want. I want the legend to show `a` in blue color and `b` in red color or vice versa — i.n.n.m, Aug 18 '17 at 21:07
the x coordinate is a and the y coordinate is b, so your legend is correct. — mauve, Aug 18 '17 at 21:08
@mauve, yes, correct, but in my legend it should show both colors and coordinates, itsn't it? — i.n.n.m, Aug 18 '17 at 21:09
@i.n.n.m, i don't understand - you have __one__ set of points with `(x,y)` coordinates, so the legend is correct... And their colors is like: `red`, 'blue', `red`, ... — MaxU - stand with Ukraine, Aug 18 '17 at 21:09
@MaxU may be I understood it wrong. Is there a way to show `a` in legend? — i.n.n.m, Aug 18 '17 at 21:13
@i.n.n.m, yes, if you want to plot __two__ columns and only if they have a __common__ `x` coordinate (for example - index of a DF)... — MaxU - stand with Ukraine, Aug 18 '17 at 21:14
@MaxU yes, I think I was a little confused, first time trying to plot a scatter plot! I wanted to plot two columns and both are indexed same — i.n.n.m, Aug 18 '17 at 21:20
@i.n.n.m, what should be the `x` axis - `df.index` or something else? — MaxU - stand with Ukraine, Aug 18 '17 at 21:21
@MaxU, I was just looking at your answer, I think that explained what I was doing. Yes, index should be `df.index` — i.n.n.m, Aug 18 '17 at 21:22

MaxU - stand with Ukraine · Accepted Answer · 2017-08-18T21:27:55.093

3

UPDATE:

plt.scatter(df.index, df.a, color='red')
plt.scatter(df.index, df.b, color='blue')
plt.legend(loc = 'best')

Result:

I think you are confused.

Demo:

Imagine you have only three rows in your DF:

In [53]: df = pd.DataFrame({'x':[1,2,3], 'y':[5,6,7]})

In [55]: df
Out[55]:
   x  y
0  1  5
1  2  6
2  3  7

You are trying to plot the following three points - what kind of legend do you expect?

In [54]: plt.scatter(df.x, df.y, color=['red','blue'])
Out[54]: <matplotlib.collections.PathCollection at 0x149f9f28>

In [56]: plt.legend(loc = 'best')
Out[56]: <matplotlib.legend.Legend at 0x1353ca20>

edited Aug 18 '17 at 21:27

answered Aug 18 '17 at 21:17

MaxU - stand with Ukraine

205,989
36
386
419

thank you for your effort to clarify , the update is the one I was looking for. – i.n.n.m Aug 18 '17 at 21:32
@i.n.n.m, glad i could help :) – MaxU - stand with Ukraine Aug 18 '17 at 21:33

Scatter plot legend shows only one variable with color -Pandas Seaborn

1 Answers1