How can I change datetime format of x-axis values in matplotlib using python?

Question

How can I make this plot's x-axis looks better?

Is there any way to show only 3 values in x-axis while plotting all values in my dataframe?

I tried using plt.xticks(rotation=70) but it still cuts my numbers..

As you can see, it shows the date in the format MMM DD YYYY .

I wanna something like MMM DD or DD MMM . Or even writing in 2 lines DD MM \\n YYYY .

How can I achieve that?

This is the code used for generating the chart:

plt.plot(df_21d["Timestamp"], df_21d["Close"], label="Price", color="b")    

plt.title("BTC - 21 Dias")

plt.xlabel("Data")
plt.ylabel("Preco (em USD)")

plt.grid(True)

plt.show()

Here's my data:

df_21d["Timestamp"]

667   2016-05-27 08:00:00
668   2016-05-27 08:30:00
669   2016-05-27 09:00:00
670   2016-05-27 09:30:00
671   2016-05-27 10:00:00
Name: Timestamp, dtype: datetime64[ns]

df_21d["Close"]

667    480.00
668    471.36
669    477.35
670    476.55
671    479.92
Name: Close, dtype: float64

I dont know if I have to format df_21d.Timestamp or if it's a matplotlib configuration. Maybe formatting the timestamp and passing it as a vector using plt.xticks() ... but I'm not sure about how do that...

Or maybe even inserting only the first, the last and the middle value for the x-axis, with something like:

plt.xaxis_labels = (df_21d["Timestamp"][1], 
             df_21d["Timestamp"][len(df_21d["Timestamp"])/2], 
             df_21d["Timestamp"][len(df_21d["Timestamp"])])

Answer 1

This answer might help you: Not write out all dates on an axis, Matplotlib

Essentially, you can set the format of dates by setting the format:

plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%d %m'))