Plot a time series grouped by id

Question

I want to plot a time series grouped by id. So that time is my x-value and 'value' is my y-value. How can I plot x and y grouped by 'id' 1?

id   time  value 
1    1     0.3
1    2     0.6
1    3     0.9
2    1     0.1
2    2     0.3
2    3     0.6
3    1     0.2
3    2     0.4
3    3     0.5

jezrael · Accepted Answer · 2016-11-21T16:05:42.880

8

I think you can use pivot with DataFrame.plot.bar or only DataFrame.plot:

import matplotlib.pyplot as plt

df = df.pivot(index='time', columns='id', values='value')
print (df)
id      1    2    3
time               
1     0.3  0.1  0.2
2     0.6  0.3  0.4
3     0.9  0.6  0.5

df.plot.bar()
plt.show()

df = df.pivot(index='time', columns='id', values='value')
print (df)
id      1    2    3
time               
1     0.3  0.1  0.2
2     0.6  0.3  0.4
3     0.9  0.6  0.5

df.plot()
plt.show()

edited Nov 21 '16 at 16:05

answered Nov 21 '16 at 15:54

jezrael

822,522
95
1,334
1,252

it's more a time series --> df.plot(subplot=True) might be better – MMF Nov 21 '16 at 15:56
@MMF - Sorry, I dont understand. Can you explain more? – jezrael Nov 21 '16 at 15:59
`I want to plot a time series grouped by id`. I think a subplot with plots from each group might be better than plotting bars to represent a time series. Just about what you're plotting actually. – MMF Nov 21 '16 at 16:00
Yes, first graph is wrong, I delete it. thank you for comment. – jezrael Nov 21 '16 at 16:01
Alright. Thanks. I what can I do if I only want to plot Id 1 ? – matthew Nov 21 '16 at 16:24
1

Then use df.loc[1].plot() – jezrael Nov 21 '16 at 16:29
1

Or Then use df.loc['1'].plot() if 1 is string value. – jezrael Nov 21 '16 at 16:30
1

Sorry, I am on phone. But you can before pivot use boolean indexing -`df=df[df.id == '1']` – jezrael Nov 21 '16 at 17:03

Plot a time series grouped by id

1 Answers1

Linked