Time Series Analysis

introml.analyticsdojo.com

59. Time Series Analysis

import numpy as np
import pandas as pd
import os
import yfinance as yf
import matplotlib.pyplot as plt
import seaborn as sns
from statsmodels.tsa.seasonal import seasonal_decompose

#!pip install yfinance

data =  yf.download(tickers="TSLA", start="2016-1-1",end="2021-12-31",progress=False)
data.head()

	Open	High	Low	Close	Adj Close	Volume
Date
2016-01-04	15.381333	15.425333	14.600000	14.894000	14.894000	102406500
2016-01-05	15.090667	15.126000	14.666667	14.895333	14.895333	47802000
2016-01-06	14.666667	14.670000	14.398667	14.602667	14.602667	56686500
2016-01-07	14.279333	14.562667	14.244667	14.376667	14.376667	53314500
2016-01-08	14.524000	14.696000	14.051333	14.066667	14.066667	54421500

plt.rcParams["figure.figsize"] = (15,10)
sns.set(font_scale=3)
data['High'].plot()
plt.xlabel('Time')
plt.ylabel('Adj Close')
plt.title('Tesla')

Text(0.5, 1.0, 'Tesla')

#Select frequency based on no of observations that comprise a cycle in the dataset
ts_decompose_add = seasonal_decompose(x=data['High'], 
                                          model='additive', 
                                          freq=180) 
estimated_trend_add = ts_decompose_add.trend
estimated_seasonal_add = ts_decompose_add.seasonal
estimated_residual_add = ts_decompose_add.resid

<ipython-input-15-ada5a077c2aa>:2: FutureWarning: the 'freq'' keyword is deprecated, use 'period' instead
  ts_decompose_add = seasonal_decompose(x=data['High'],

fig, axes = plt.subplots(4, 1, sharex=True, sharey=False)
# fig.set_figheight(10)
# fig.set_figwidth(15)

axes[0].plot(data['High'], label='Original')
axes[0].legend(loc='upper left');

axes[1].plot(estimated_trend_add, label='Trend')
axes[1].legend(loc='upper left');

axes[2].plot(estimated_seasonal_add, label='Seasonal')
axes[2].legend(loc='upper left');

axes[3].plot(estimated_residual_add, label='Residuals')
axes[3].legend(loc='upper left');