CODEWITHSUNDEEP
python-2.7timematplotlibpandastime-series
Jason
Jason

Reputation: 4546

Unexpected plotting behaviour using matplotlib and time series data

I'd like to plot a line graph of a time series using matlplotlib, however matplotlib will not plot all of my data unless I use .plot_date() or use .plot() passing o as the argument for the markers.

If I use .plot_date or .plot() and using the '-' marker, my data doesn't plot correctly:

enter image description here

Does anybody know why this is happening and how it can be fixed? I need the data points to be connect with lines.

Thanks in advance.

Here's my current code:

import pandas as pd
import matplotlib.pyplot as plt
import dateutil
from datetime import date, datetime, timedelta

plt.plot(sve2_all['MeHg ng/l']['1993-01-18':'1997-05-02'].index, sve2_all['MeHg ng/l']['1993-01-18':'1997-05-02'],'bo')
plt.xticks(rotation=70)
plt.show()

My data is in a Pandas DataFrame and the index is datetime64.

Upvotes: 2

Views: 1215

Answers (2)

Nasser Al-Wohaibi
Nasser Al-Wohaibi

Reputation: 4661

For Pandas's Series you can just ser.dropna() as @RutgerKassies suggested. However this will not work for DataFrames. For DataFrames someone solved the issue using masks. Or you can just df.fillna(method='ffill'):

sve2_all['MeHg ng/l ffill']= sve2_all['MeHg ng/l'].fillna(method='ffill')
sve2_all['MeHg ng/2 ffill']= sve2_all['MeHg ng/2'].fillna(method='ffill')
sve2_all[['MeHg ng/l ffill','MeHg ng/2 ffill']].plot()

make sure forward filling is logical to your data model.

Upvotes: 0

Rutger Kassies
Rutger Kassies

Reputation: 64463

You probably have a lot of NaN values in your Dataframe. Matplotlib only draws a line between consecutive (valid) data points, and leaves a gap at NaN values.

If that's the case, removing the NaN's before plotting should do the trick. For example:

dftmp = sve2_all['MeHg ng/l']['1993-01-18':'1997-05-02'].dropna()
plt.plot(dftmp.index, dftmp,'b-')

Upvotes: 2

Related Questions