autocorrelation_spectrum_python

"""
If it will be helpful, please like this post
"""

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import scipy.stats
from statsmodels.graphics.tsaplots import plot_acf
#https://scicoding.com/4-ways-of-calculating-autocorrelation-in-python/

if __name__ == '__main__':
    df = pd.read_csv('CompanyABCProfit.csv').set_index('Year')
    # fig = plt.figure(figsize=(10, 10))
    # plt.xlabel('Year')
    # plt.ylabel('Profit')
    # df.plot()
    # plt.close()
    values_df = []
    for value in df["Profit(Rs '000)"]:
        values_df.append(value)
    corr_func = (np.correlate(values_df - df.mean()[0], values_df - df.mean()[0], 'full')) / df.var()[0] / len(values_df)
    # fig = plt.figure(figsize=(10, 10))
    corr_function = pd.DataFrame(corr_func).reset_index()
    corr_function['index'] = corr_function['index'] - len(df)
    corr_function = corr_function.set_index('index')
    corr_function.plot()
    # plt.show()
    # plt.close()
    Fs = 1024
    sampling_rate = 30 #signal freq
    Number_of_repeat = 10
    N = int(Number_of_repeat * Fs / sampling_rate) #number of samples
    t_step = 1 / Fs
    t = np.linspace(0, (N - 1) * t_step, N) #time interval
    freq = Fs / N #freq step
    f = np.linspace(0, (N - 1) * freq, N) # freq interval

    data = 2 * np.sin(2 * np.pi * sampling_rate * t) + 4 * np.sin(2 * np.pi * 3 * sampling_rate * t)
    fureir = scipy.fft.fft(data)
    fureier_abs = abs(fureir / N)
    data_spec = f[0 : int(N / 2 + 1)]
    furier_abs_plot = 2 * fureier_abs[0 : int(N / 2 + 1)]
    fureier_abs[0] /= 2
    fig, [ax1, ax2] = plt.subplots(nrows=2, ncols=1)
    ax1.plot(t, data)
    ax2.plot(data_spec, furier_abs_plot)
    plt.show()
    stop = 'here'