Matplotlib key error with dataframe

Question

I am trying to display some points using matplotlib, Although I can display them using print command but matplotlib gives error. The command that works is also there(commented).

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline

data = np.array([[-1,-1,'C1'],[-2,-1,'C1'],[-3,-2,'C1'],[1,1,'C2'],[2,1,'C2'],[3,2,'C2']])
query=[-2.5,-1.5]

df=pd.DataFrame(data)
df.columns =['x','y','Cat']
df

for i in range(6):
    if(df.ix[i]['Cat'] == 'C1'):
        plt.scatter(df.iloc[i]['x'], df.iloc[i]['y'], s=150, c='r') #error line
         #working linke below
         #print(df.iloc[i]['x'],df.iloc[i]['y'])
    else:
        plt.scatter(df.iloc[i]['x'], df.iloc[i]['y'], s=150, c='b')
        #working line below
        #print(df.iloc[i]['x'],df.iloc[i]['y'])

Please help. Thanks in advance

Thanks @Haleemur Ali for your help I am able to run it now but still not fully functional. Not all points are showing not sure why?

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline

data = np.array([[-1,-1,'r'],[-2,-1,'r'],[-3,-2,'r'],[1,1,'b'],[2,1,'b'],[3,2,'b'],[-2.5,-1.5,'y']])
query=[-2.5,-1.5]

df=pd.DataFrame(data)
df.columns =['x','y','Cat']
print(df)

plt.scatter(df.x, df.y, s=150, c=df.Cat)

Graph generated

enter image description here

In contrast to builtin list numpy arrays can only hold data of the same type. In your case data is cast to a string type (<u21) because of the strings on the second index. Matplotlib does not know how to plot non-numeric data; that's the problem here. — MaxPowers
– MaxPowers, Commented Mar 2, 2018 at 17:25

ImportanceOfBeingErnest · Accepted Answer · 2018-03-02 17:49:34Z

2

If the numbers are strings, they are not recognized as numbers, hence they are plotted as categories, just as you would expect if you plotted ["apple", "banana", "cherry"]. You would need to convert your data to floats:

df[['x', 'y']] = df[['x', 'y']].astype(float)

Complete code:

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

data = np.array([[-1,-1,'r'],[-2,-1,'r'],[-3,-2,'r'],[1,1,'b'],
                 [2,1,'b'],[3,2,'b'],[-2.5,-1.5,'y']])

df=pd.DataFrame(data, columns=['x','y','Cat'])
df[['x', 'y']] = df[['x', 'y']].astype(float)

plt.scatter(df.x, df.y, s=150, c=df.Cat)

plt.show()

answered Mar 2, 2018 at 17:49

ImportanceOfBeingErnest

342k61 gold badges737 silver badges771 bronze badges

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

3 Comments

Haleemur Ali Over a year ago

this is true! What I find interesting is that matplotlib correctly inferred that the series passed to scatter were actually numeric, as you can see in my answer, but clearly not the case for the OP. BTW, I'm using the following versions on Windows Anaconda pandas=0.20.3, matplotlib=2.0.2. What versions are you using?

ImportanceOfBeingErnest Over a year ago

In matplotlib < 2.1, strings are converted to floats if possible, else an error is raised. In matplotlib >= 2.1 strings are interpreted as categorical values.

Nitin Over a year ago

I am using 2.1.0 version of matplotlib. That clears the confusion. But thing is even if matplotlib interpreted values as category why it did not displayed all the data. Thanks – ImportanceOfBeingErnest & – Haleemur Ali.

Haleemur Ali · Accepted Answer · 2018-03-02 15:49:56Z

1

Scatter plots aren't built by iterating through the data.

You can build the scatter plot for a particular category, like this:

plt.scatter(df.x[df.Cat=='C1'], df.y[df.Cat=='C1'], s=150, c='r')

You can also create a scatter plot where each category gets a distinct colour

plt.scatter(df.x, df.y, s=150, c=df.Cat)

edited Mar 2, 2018 at 15:49

answered Mar 2, 2018 at 15:43

Haleemur Ali

28.6k6 gold badges67 silver badges89 bronze badges

Collectives™ on Stack Overflow

Matplotlib key error with dataframe

2 Answers 2

3 Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related