CODEWITHSUNDEEP
pythonpandasstatsmodelsarima
Python coder
Python coder

Reputation: 849

MatPlotlib: X-axis start from year 1970 but is from only 2017

I have dataframe with timestamp as index and price values as column. When I try to plot using plot_acf, the x-axis is starting from 1970.

Code:

import pandas as pd
from statsmodels.graphics.tsaplots import plot_acf, plot_pacf

data = {'price_btc': {Timestamp('2017-04-04 00:00:00'): 1132.0,
               Timestamp('2017-04-05 00:00:00'): 1142.0,
               Timestamp('2017-04-06 00:00:00'): 1128.0,
               Timestamp('2017-04-07 00:00:00'): 1164.0,
               Timestamp('2017-04-08 00:00:00'): 1189.0,
               Timestamp('2017-04-09 00:00:00'): 1188.0,
               Timestamp('2017-04-10 00:00:00'): 1194.0,
               Timestamp('2017-04-11 00:00:00'): 1208.0,
               Timestamp('2017-04-12 00:00:00'): 1213.0,
               Timestamp('2017-04-13 00:00:00'): 1218.0}}

df = pd.DataFrame(data)

# Original Series
fig, axes = plt.subplots(3, 2, sharex=True, figsize=(20, 5))
axes[0, 0].plot(df.price_btc); axes[0, 0].set_title('Original Series')
plot_acf(df.price_btc, ax=axes[0, 1])

# 1st Differencing
axes[1, 0].plot(df.price_btc.diff()); axes[1, 0].set_title('1st Order Differencing')
plot_acf(df.price_btc.diff().dropna(), ax=axes[1, 1])

# 2nd Differencing
axes[2, 0].plot(df.price_btc.diff().diff()); axes[2, 0].set_title('2nd Order Differencing')
plot_acf(df.price_btc.diff().diff().dropna(), ax=axes[2, 1])

plt.show()

Output: enter image description here

Expected output is to have dates from 2017 in Autocorrelation plots as well.

Issue is with sharex=True. Need to have different x lables for autocorrelation plot.

Upvotes: 2

Views: 1705

Answers (1)

Arne Decker
Arne Decker

Reputation: 923

First of all, you do not need .diff().diff() for second order differencing. You can give a integer as order to the .diff() method, so second order would be .diff(2).

If you set sharex = True then all x-axis will be the same. The autocorrelation plot is not scaled by dates or so but by lags. Set sharex = False and add this line here for each plot to set the x axis:

axes[0, 0].set_xlim(dt.datetime(2017,4,4), dt.datetime(2017,4,13))

You can set the start and end date then. Here is the full code:

import pandas as pd
from statsmodels.graphics.tsaplots import plot_acf, plot_pacf
import matplotlib.pyplot as plt
import datetime as dt
import matplotlib.dates as mdates

data = {'price_btc': {pd.Timestamp('2017-04-04 00:00:00'): 1132.0,
           pd.Timestamp('2017-04-05 00:00:00'): 1142.0,
           pd.Timestamp('2017-04-06 00:00:00'): 1128.0,
           pd.Timestamp('2017-04-07 00:00:00'): 1164.0,
           pd.Timestamp('2017-04-08 00:00:00'): 1189.0,
           pd.Timestamp('2017-04-09 00:00:00'): 1188.0,
           pd.Timestamp('2017-04-10 00:00:00'): 1194.0,
           pd.Timestamp('2017-04-11 00:00:00'): 1208.0,
           pd.Timestamp('2017-04-12 00:00:00'): 1213.0,
           pd.Timestamp('2017-04-13 00:00:00'): 1218.0}}

df = pd.DataFrame(data)

# Original Series
fig, axes = plt.subplots(3, 2, sharex=False, figsize=(20, 5))
axes[0, 0].plot(df.price_btc); axes[0, 0].set_title('Original Series')
axes[0, 0].set_xlim(dt.datetime(2017,4,4), dt.datetime(2017,4,13))
plot_acf(df.price_btc, ax=axes[0, 1])

# 1st Differencing
axes[1, 0].plot(df.price_btc.diff()); axes[1, 0].set_title('1st Order     Differencing')
axes[1, 0].set_xlim(dt.datetime(2017,4,4), dt.datetime(2017,4,13))
plot_acf(df.price_btc.diff(1).dropna(), ax=axes[1, 1])

# 2nd Differencing
axes[2, 0].plot(df.price_btc.diff().diff()); axes[2, 0].set_title('2nd Order Differencing')
axes[2, 0].set_xlim(dt.datetime(2017,4,4), dt.datetime(2017,4,13))
plot_acf(df.price_btc.diff(2).dropna(), ax=axes[2, 1])

plt.show()

Upvotes: 1

Related Questions