如何减少在matplotlib图上显示的xticks数量？

Question

I'm plotting time series data and I would like to maintain the resolution in the curves, but I want to display fewer years on the x-axis. I've parsed the 'Year' column as dates and I've made that column into the dataframe's index. I feel like it should be easy to thin down the frequency of the labels, but everything I've tried has simply made the labels inaccurate.

The DataFrame has the form:

print(total_df.head())

         All ages  Age 18 or older
Year                           
1978     131.0            183.0
1979     133.0            185.0
1980     138.0            191.0
1981     153.0            211.0
1982     170.0            232.0

And I've been using this code to produce my plot.

with sns.axes_style("whitegrid"):
    fig, ax = plt.subplots(figsize=(10,7))
    ax.plot(total_df['All ages'])
    ax.plot(total_df['Age 18 or older'])
    ax.set_title('Total Imprisonment Rates (table: p16f01)')
    ax.set_xlabel('Year')
    ax.set_ylabel('People imprisoned (per 100k US population)')
    ax.set_xticklabels(total_df.index, rotation=70)
    ax.legend()
    ax.set_ylim([0, 1.1*max([total_df['All ages'].max(), 
                             total_df['Age 18 or older'].max()])])

Which produces

Answer 1

This shouldn't happen. And it wouldn't happen if you made sure your index are actual numbers, not strings.

To convert your index to numbers, use eg

df.index = df.index.values.astype(int)

Then remove the set_xticklabels line, because this would anyway only make sense if you set the ticks via set_ticks as well. This will then assure that matplotlib automatically chooses useful spacings between the ticks.

Complete example:

import matplotlib.pyplot as plt
import numpy as np;np.random.seed(9)
import pandas as pd

inx = np.arange(1978,2017).astype(str)
a = np.cumsum(np.random.randn(len(inx),2), axis=0)+10
df = pd.DataFrame(a, index=inx, columns=list("AB"))

df.index = df.index.values.astype(int)

fig, ax = plt.subplots(figsize=(10,7))
ax.plot(df['A'])
ax.plot(df['B'])

ax.set_xlabel('YEAR')

ax.legend()
ax.set_ylim([0, 1.1*max([df['A'].max(), 
                         df['B'].max()])])

plt.show()