Dipping my toe into Time Series Analysis

2021-Nov-18, Thursday

I found a tutorial on time series analsyis (TSA), which uses the Python library statsmodels to decompose a time series (TS). I thought it'll be fun to play around with it a bit.
Here's the link to the tutorial: https://towardsdatascience.com/time-series-decomposition-in-python-8acac385a5b2

They use TS data from: https://www.kaggle.com/rakannimer/air-passengershttps://www.kaggle.com/rakannimer/air-passengers

• In TSA it is useful to split a TS into its different components representing trend, seasonality, and noise.

• Trend: movement of time series over time; can be increasing, decreasing, or constant.

• Seasonality: periodic part of the time series.

• Noise: what remains behind after removing trend and seasonality. This is the variability in the time series that cannot be explained by the model.

import numpy as np
import pandas as pd
from statsmodels.tsa.seasonal import seasonal_decompose

df = pd.read_csv("AirPassengers.csv")

df.head()

	Month	#Passengers
0	1949-01	112
1	1949-02	118
2	1949-03	132
3	1949-04	129
4	1949-05	121

# Set the "Month" column as index, and convert format to Datetime:
df.set_index("Month", inplace=True)
df.index = pd.to_datetime(df.index)

df.head()

	#Passengers
Month
1949-01-01	112
1949-02-01	118
1949-03-01	132
1949-04-01	129
1949-05-01	121

# Check for null values:
df.isnull().sum()

#Passengers    0
dtype: int64

df.plot()

<AxesSubplot:xlabel='Month'>

ts_decomposed = seasonal_decompose(df["#Passengers"], model="multiplicative", period=12)

ts_decomposed.seasonal.plot()

<AxesSubplot:xlabel='Month'>

ts_decomposed.trend.plot()

<AxesSubplot:xlabel='Month'>

ts_decomposed.resid.plot()

<AxesSubplot:xlabel='Month'>