How can I keep the data from being crowded together in matplotlib scatter plot?

Question

I run my code, and then I get the following figure.

When I modify the value of figure options I get the output I want.

What should I do with my code? How can I set the left and right values in code?

from matplotlib import pyplot as plt
import matplotlib.dates as mdate
import pandas as pd 
import numpy as np
import datetime
from decimal import Decimal
from pandas.plotting import register_matplotlib_converters
register_matplotlib_converters()

data =[(datetime.date(2019, 12, 23), Decimal('0.3230'), Decimal('157.89')),
       (datetime.date(2019, 12, 23), Decimal('0.1680'), Decimal('157.74')),  
       (datetime.date(2019, 12, 23), Decimal('0.1440'), Decimal('152.78')),  
       (datetime.date(2019, 12, 24), Decimal('0.1070'), Decimal('157.94')),  
       (datetime.date(2019, 12, 24), Decimal('0.6420'), Decimal('159.81')),   
       (datetime.date(2019, 12, 25), Decimal('0.0980'), Decimal('115.31')),   
       (datetime.date(2019, 12, 26), Decimal('0.0830'), Decimal('122.89')),    
       (datetime.date(2019, 12, 26), Decimal('0.1600'), Decimal('171.88')),   
       (datetime.date(2019, 12, 27), Decimal('0.1710'), Decimal('161.40')),   
       (datetime.date(2019, 12, 30), Decimal('0.0590'), Decimal('161.02')),   
       (datetime.date(2019, 12, 30), Decimal('0.4150'), Decimal('168.19')),   
       (datetime.date(2019, 12, 31), Decimal('0.1150'), Decimal('167.83')),   
       (datetime.date(2020, 1, 2), Decimal('0.0790'), Decimal('192.41')),   
       (datetime.date(2020, 1, 3), Decimal('0.4930'), Decimal('216.43')),   
       (datetime.date(2020, 1, 3), Decimal('0.2640'), Decimal('154.92')),   
       (datetime.date(2020, 1, 3), Decimal('0.1980'), Decimal('211.11')),   
       (datetime.date(2020, 1, 10), Decimal('0.2173'), Decimal('248.64')),   
       (datetime.date(2020, 1, 13), Decimal('1.3202'), Decimal('196.34')),   
       (datetime.date(2020, 1, 14), Decimal('0.0423'), Decimal('198.30')),   
       (datetime.date(2020, 1, 16), Decimal('0.0236'), Decimal('296.56')),   
       (datetime.date(2020, 1, 16), Decimal('0.0937'), Decimal('304.03'))]

df = pd.DataFrame(data)
df.rename(columns={0:'a',1:'b',2:'c'},inplace=True)
x = df.a
y = df.c
fig,ax=plt.subplots()
ax.scatter(x,y,facecolor='g',alpha=0.5)
ax.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m-%d'))
ax.set_xticks(x.drop_duplicates())
fig.autofmt_xdate()
plt.show()

Answer 1

You need to set the x-axis limits directly using ax.set_xlim . You can get the maximum and minimum from the dataframe returned by x.drop_duplicates() and then pad by a small time delta like this

# ...
_x = x.drop_duplicates()
ax.set_xticks(_x)
# Create a 2-day timedelta and set the limits from the min and max with td as padding
td = datetime.timedelta(days=2)
ax.set_xlim([min(_x)-td, max(_x)+td])
fig.autofmt_xdate()
plt.show()

This will give you

I would recommend also setting a wider figure, ie

fig, ax = plt.subplots(figsize=(10, 5))

Which will give you

Answer 2

This is an unfortunate consequence of autoscaling of collections. The issue is #7413 and a solutions is in #13642 . The problem will hence not occur in matplotlib 3.2.

Up to then, you can do a hacky manual autoscaling by adding and removing a line plot of the same data as the scatter. Ie in the code from the question, replace ax.scatter(...) by

def myautoscale(sc):
    ax = sc.axes
    x, y= sc.get_offsets().T
    line, = ax.plot(x,y)
    ax.relim()
    ax.autoscale_view()
    line.remove()

sc = ax.scatter(x,y, facecolor='g', alpha=0.5)
myautoscale(sc)