Autocorrelation plots are specifically used to check the randomness between data points of a dataset. In time series analysis, correlation is computed parallel to each data point at varying time lags to determine their relationship.
pandas.plotting.autocorrelation_plot()
pandas.plotting.autocorrelation_plot(series, ax=None, **kwargs)
series: It should be a time series instance.ax: It shows matplotlib axis object. Its default value is None.**kwargs: These are keyword arguments.It returns a matplotlib.axis.Axes object.
In this example, we draw an autocorrelation plot on randomly generated data points.
# importing librariesimport pandas as pdimport matplotlib.pyplot as plotimport numpy as np# creating a sample space of 500 valuesdata = np.linspace(-10, np.pi*5, num=500)# creating a series of random values_series = pd.Series(np.cos(data) * np.random.rand(500))# generate autocorrelation plotpd.plotting.autocorrelation_plot(_series)# save above generated graph as PNG file in output directoryplot.savefig("output/graph.png")
pandas, matplotlib, and numpy libraries.np.linspace() to generate 500 sample values between -10 and np.pi * 5. _series series of random numbers.pd.plotting.autocorrelation_plot() method generates an autocorrelation plot of the above-created series of random numbers.