how to perform conditional area plotting with matplotlib?

Question

I have created the following dataframe based on a range of data.

df['data_classification'] = df.myDatarange.apply(lambda a:'Very good' if a>=-90 
                                       else ('Good' if (a>= -100 or a<=-91) 
                                             else ('Moderate' if (a>= -110 or a<=-101) 
                                                   else ('Poor' if (a>= -123 or a<=-111) 
                                                         else ('Bad' if (a>= -140 or a<=-124) 
                                                               else 'Off' )))))

I am planning to plot myDatarange with data_classification and somehow show the relation with different colour. I am very confused how to plot this. I can plot myDatarange as a single lineplot, but how to relate the two data?

So far, I have tried the following:

x1 = df1.index
y1 = df1.myDatarange
f, (ax1,ax2) = plt.subplots(2,figsize=(5, 5))
ax1.plot(x1,y1,color='red', linewidth=1.9, alpha=0.9, label="myDataRange")
plt.show()

How can I plot the above range of data based on classification as area plot? Is there a better way than area plot to express my data? There are examples on the net, but not very clear on conditional side of it.

Look at using pd.cut instead of if-else structure for neater more readable code in my opinion. — Scott Boston
– Scott Boston, Commented Jun 14, 2020 at 13:59

JohanC · Accepted Answer · 2020-06-14 18:31:41Z

2

Seaborn's barplot can take a hue parameter to color each bar corresponding to the 'data_classification'. The new 'data_classification' column can be created quicker and easier to modify via pd.cut.

The barplot can be used as background for the lineplot to show the classification of each value.

Here is an example to get you started:

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

df = pd.DataFrame({'myDatarange': np.random.randint(-150, -50, size=50)})
ranges = [-10**6, -140, -123, -110, -100, -90, 10**6]
df['data_classification'] = pd.cut(df['myDatarange'], ranges, right=False,
                                   labels=['Off', 'Bad', 'Poor', 'Moderate', 'Good', 'Very Good'])

fig, ax1 = plt.subplots(figsize=(12, 4))
ax1.plot(df.index, df['myDatarange'], color='blue', linewidth=2, alpha=0.9, label="myDataRange")

sns.barplot(x=df.index, y=[df['myDatarange'].min()] * len(df),
            hue='data_classification', alpha=0.5, palette='inferno', dodge=False, data=df, ax=ax1)
for bar in ax1.patches: # optionally set the bars to fill the complete background, default seaborn sets the width to about 80%
    bar.set_width(1)

plt.legend(bbox_to_anchor=(1.02, 1.05) , loc='upper left')
plt.tight_layout()
plt.show()

PS: If you want to the 0 at the bottom (now at the top due to the negative y-values), you could call ax.invert_yaxis().

edited Jun 14, 2020 at 18:31

answered Jun 14, 2020 at 14:36

JohanC

81.4k8 gold badges54 silver badges90 bronze badges

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

4 Comments

codeheadache Over a year ago

Hello JohanC and Scot.Thank you for your kind responses. I tried Johns solution.It helped me clear some of the concepts but not what i intend to show as output. Is it possible to simply have two plots, one myDatarange line plot. Below that plot, another plot of some kind which simple shows one of the mentioned categories ie good,bad etc? I also thought of simply using a different colour for different category of myDatarange. But i don't understand the concept enough to be able to implement it.

JohanC Over a year ago

Could you edit your original question and add more information, maybe some simple data as example and a desired plot (possibly with annotations using a painting software)? What exactly do you mean by "area plot"? Why do you want two subplots when all the information fits in one? (You can just call above code and change ax=ax2 if you want the two parts separated).

codeheadache Over a year ago

JohanC. The recent edit you made answers my question. Thank you for your support

codeheadache Over a year ago

Sorry i could not do that before, because my points were very less to upvote. now i can

Collectives™ on Stack Overflow

how to perform conditional area plotting with matplotlib?

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related