20
import pandas as pd
import matplotlib.pyplot as plt

file = 'd:\\a\\pandas\\test.xlsx'
data = pd.ExcelFile(file)
df1 = data.parse('Link')
df2 = df1[['dataFor', 'total']]
df2

returns:

enter image description here

 print (type(df2))

tells me

class 'pandas.core.frame.DataFrame'

trying

df2.plot(kind='line')

returns

matplotlib.axes._subplots.AxesSubplot at 0xe4241d0

Could it be the environment?

Jupyter notebook > Help > About

The version of the notebook server is 4.2.3 and is running on:
Python 3.5.2 |Anaconda 4.2.0 (32-bit)| (default, Jul  5 2016, 11:45:57) [MSC    v.1900 32 bit (Intel)]

Where is the fault? Is matplotlib still the standard or should beginners go for Bokeh or for both?

Ryan
  • 2,167
  • 2
  • 28
  • 33
vbd
  • 3,437
  • 4
  • 32
  • 45

3 Answers3

51

In case you want to see the plot inline, use

%matplotlib inline

in the header (before the imports).

If you want to show the graphic in a window, add the line 

plt.show()

at the end (make sure you have imported import matplotlib.pyplot as plt in the header).

ImportanceOfBeingErnest
  • 321,279
  • 53
  • 665
  • 712
  • 1
    missed plt.show() ... note to myself never post on friday late afternoon... – vbd Jan 27 '17 at 16:49
  • 2
    for anyone else looking for why [missingno](https://github.com/ResidentMario/missingno) library is not displaying plots in a jupyter, use `%matplotlib inline` in single cell. (figured it out thanks to this answer) – stormfield Feb 10 '20 at 16:06
  • %matplotlib inline place before the res.plot() not before import in jupyter. – temple Nov 16 '20 at 08:29
4
## importing libraries
## notice to import %matplotlib inline to plot within notebook
import pandas as pd
%matplotlib inline
import matplotlib.pyplot as plt
import datetime


## making a DF like yours
df2 = pd.DataFrame([], columns=['dataFor','total'])
df2['dataFor'] = [datetime.datetime(2013, 9, 11),datetime.datetime(2013, 9, 12),datetime.datetime(2013, 9, 13),datetime.datetime(2013, 9, 14),datetime.datetime(2013, 9, 15),datetime.datetime(2013, 9, 16),datetime.datetime(2013, 9, 17)]
df2['total'] = [11,15,17,18,19,20,21]

## notice date are datetimes objects and not strings
df2.plot(kind='line')

output:

enter image description here

if one wants to improve graph layout:

plt.figure(figsize=(20,10))
plt.plot(df2.dataFor, df2.total, linewidth=5)
plt.plot(df2.dataFor, df2.total, '*', markersize=20, color='red')
plt.xticks(fontsize=20, fontweight='bold',rotation=90)
plt.yticks(fontsize=20, fontweight='bold')
plt.xlabel('Dates',fontsize=20, fontweight='bold')
plt.ylabel('Total Count',fontsize=20, fontweight='bold')
plt.title('Counts per time',fontsize=20, fontweight='bold')
plt.tight_layout()

enter image description here

epattaro
  • 2,330
  • 1
  • 16
  • 29
0

I got the same problem on performing the plot but I got it solved by running the import clause followed by %matplotlib. I think I'm using the latest version. I've tried the " %matplotlib inline " but for some reasons, it doesn't work on my end.

Qi Wang
  • 1