Unable to add lines to `axes` used by pandas `plot` method

Question

I find the pandas DataFrame.plot method very useful and I especially like the way it formats the x ticks when the x variable is a date. However, there are times when I want to add more complicated marks on top of a plot that was generated using DataFrame.plot. For example, plotting a time series using ax = my_timeseries_df.plot("date", "value) and then calling the fill_between method on the ax object to add confidence intervals. However, when I call a plotting method on an ax object after using it with pandas I can't get any marks to show up.

Here is an example:

And here is the raw code:

import matplotlib.pyplot as plt
import pandas as pd

test = pd.DataFrame({
    "a": pd.date_range(start="2020-04-21", periods=5, freq="D"),
    "b": [1, 2, 3, 4, 5],
    "c": [2, 4, 6, 8, 10],
})
fig, ax = plt.subplots()
test.plot("a", "b", ax=ax)
ax.plot(test.a, test.c, color="red", linewidth=20)

Strangely the y-limit changes after calling ax.plot(...) to accommodate the new points, but the new line does not appear.

I have tried adding %matplotlib inline to the cell but that doesn't help. I could always just bypass DataFrame.plot and use matplotlib for everything, but dealing with time-based x-ticks directly in matplotlib is annoying enough that I would love to let pandas do it for me.

Please copy and paste the example code into your answer, so we can copy and paste it into our editors. — Joooeey
– Joooeey, Commented Jun 30, 2020 at 20:43
This question may be relevant: stackoverflow.com/questions/29329725/… — Joooeey
– Joooeey, Commented Jun 30, 2020 at 20:45
You could try ax.set_xlim(auto=True) to figure out if the line is plotted on a different part of the x-axis. This can happen when using different modules (Pandas and Matplotlib) that handle the dates differently. — Joooeey
– Joooeey, Commented Jun 30, 2020 at 20:47

Scott Boston · Accepted Answer · 2020-06-30 21:05:33Z

1

Use x_compat=True, see docs here:

import matplotlib.pyplot as plt
import pandas as pd

test = pd.DataFrame({
    "a": pd.date_range(start="2020-04-21", periods=5, freq="D"),
    "b": [1, 2, 3, 4, 5],
    "c": [2, 4, 6, 8, 10],
})
fig, ax = plt.subplots()
test.plot("a", "b", ax=ax, x_compat=True)
ax.plot(test.a, test.c, color="red", linewidth=20)

Output:

answered Jun 30, 2020 at 21:05

Scott Boston

154k15 gold badges160 silver badges207 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Max Taggart Over a year ago

You are correct. It's a bummer that you lose the tick formatting though. The formatting is pretty much the only reason I use the pandas plot instead of just doing it all in matplotlib.

Max Taggart · Accepted Answer · 2020-06-30 21:35:33Z

Also, for reference, matplotlib has a couple of classes that make it much easier to get closer to the tick formatting that pandas has baked in; AutoDateLocator and ConciseDateFormatter. Although they still don't get you complete parity. You can get the docs here.

import matplotlib.dates as mdates
import matplotlib.pyplot as plt
import pandas as pd

test = pd.DataFrame({
    "a": pd.date_range(start="2020-04-21", periods=5, freq="D"),
    "b": [1, 2, 3, 4, 5],
    "c": [2, 4, 6, 8, 10],
})
fig, ax = plt.subplots()
locator = mdates.AutoDateLocator(minticks=3, maxticks=10)
formatter = mdates.ConciseDateFormatter(locator)
ax.xaxis.set_major_locator(locator)
ax.xaxis.set_major_formatter(formatter)

ax.plot(test.a, test.b)
ax.plot(test.a, test.c, color="red", linewidth=2)

Collectives™ on Stack Overflow

Unable to add lines to `axes` used by pandas `plot` method

2 Answers 2

1 Comment

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related