CODEWITHSUNDEEP
pythonpandasmatplotlib
AGH_TORN
AGH_TORN

Reputation: 833

How Can I Change X Labels In Pandas Plot?

I've got an Excel Worksheet that has columns of specific statuses, I need a count of each status and then I need it to be graphed in a bar plot. Example:

Status_1    Status_2    Status_3
Active      Abandoned   Active
Inactive    Abandoned   Active

Currently I've been using the following code:

import pandas as pd, matplotlib.pyplot as plt
report = r"\\Myserver\Reports\Report.xlsx"
df1 = pd.read_excel(report, sheetname=1)
df = df1[['Status_1', 'Status_2', 'Status_3']].copy()

val1 = df['Status_2'].value_counts().to_frame()
vals1 = pd.DataFrame(val1)
fig, axes = plt.subplots(nrows=2, ncols=2)
vals1.plot(kind='bar', ax=axes[0,0])

So my problem now is that the graph displays properly, but the columns on the X axis are labeled with their statuses (Active/Inactive) and the legend shows the column name (Status_1). I'd like this to be reversed and have the X axis only show "Status_1" and the legend to be color coded with two choices of "Active" and "Inactive".

Because I'm doing this with value counts it's gotten be very confused and turned around. Is there any way of doing this?

enter image description here

Upvotes: 1

Views: 3116

Answers (2)

jrjc
jrjc

Reputation: 21873

You can use seaborn, along with reshaping you dataframe into a tidy form 

import seaborn as sns
import pandas as pd

In [26]: df
Out[26]: 
    Status_1   Status_2   Status_3
0     Active  Abandoned     Active
1   Inactive  Abandoned  Abandoned
2     Active  Abandoned     Active
3   Inactive   Inactive     Active
4     Active  Abandoned   Inactive
5  Abandoned  Abandoned     Active

In [27]: df2 = pd.melt(df) # convert into tidy data
In [28]: df2
Out[28]: 
    variable      value
0   Status_1     Active
1   Status_1   Inactive
2   Status_1     Active
3   Status_1   Inactive
4   Status_1     Active
[...]
15  Status_3     Active
16  Status_3   Inactive
17  Status_3     Active
In [29]: sns.factorplot(data=df2, x="variable", kind="count", hue="value")

factorplot

Or:

sns.factorplot(data=df2, x="variable", kind="count", col="value")

factoroplot

HTH

Upvotes: 2

Kacper Wolkowski
Kacper Wolkowski

Reputation: 1597

I hope this will help:

a = pd.DataFrame([['a','i'],['a','i'],['i','a']], columns=["S1", "S2"])
a.apply(lambda x: x.value_counts()).transpose().plot.bar()

a looks like this

Result:

enter image description here

Explanation:

a.apply(lambda x: x.value_counts()) - creates a dataframe of counted values from your data,

.transpose() - transpose index and columns (in case if you want to go back to having Status1, Status2 as your bars and Active/Inactive as x-label, just remove this method,

plot.bar() - just plots your DataFrame

Upvotes: 1

Related Questions