CODEWITHSUNDEEP
pythonmatplotlibbar-chartseaborn
Jayashree
Jayashree

Reputation: 825

Seaborn barplot ordering by bar length

I am trying to plot barplot using seaborn.

k= 'all'
k_best = SelectKBest(k=k)
k_best=k_best.fit(features, labels)
features_k=k_best.transform(features)
scores = k_best.scores_ # extract scores attribute
pairs = zip(features_list[1:], scores) # zip with features_list
pairs= sorted(pairs, key=lambda x: x[1], reverse= True) # sort tuples in descending order
print pairs

#Bar plot of features and its scores
sns.set(style="white")
ax = sns.barplot(x=features_list[1:], y=scores)
plt.ylabel('SelectKBest Feature Scores')
plt.xticks(rotation=90)

My plot looks like this enter image description here

I want the bars to be sorted in descending order.Exercised_stock_options with highest value on left followed by total_stock_value and so on.

Kindly help.Thanks

Upvotes: 3

Views: 4350

Answers (1)

ImportanceOfBeingErnest
ImportanceOfBeingErnest

Reputation: 339250

The two lines

pairs = zip(features_list[1:], scores) # zip with features_list
pairs= sorted(pairs, key=lambda x: x[1], reverse= True)

already give you a list of tuples, sorted by the scores values. Now you only need to unzip it to two lists, which can be plotted.

newx, newy = zip(*pairs)
sns.barplot(x=newx, y=newy)

A full working example:

import seaborn.apionly as sns
import matplotlib.pyplot as plt

x = ["z","g","o"]
y = [5,7,4]

pairs = zip(x, y)
pairs= sorted(pairs, key=lambda x: x[1], reverse= True)

newx, newy = zip(*pairs)
ax = sns.barplot(x=newx, y=newy)

plt.show()

enter image description here

Upvotes: 2

Related Questions