Time Series Analysis Basics

Introduction

Time series analysis is used to study data points collected over time. It helps in identifying trends, seasonal variations, and patterns in data. It is widely used in fields like finance, stock market analysis, weather forecasting, sales forecasting, and more.

In this tutorial, we will cover:

1. What is Time Series Data?
2. Working with Time Series Data in Pandas
3. Resampling and Aggregation
4. Rolling Statistics (Moving Average)
5. Detecting Trends and Seasonality
6. Time Series Forecasting Basics

1. What is Time Series Data?

Time series data is a sequence of observations recorded at time intervals (e.g., hourly, daily, weekly). Examples include:

• Stock prices recorded daily
• Monthly sales of a store
• Temperature readings every hour

2. Working with Time Series Data in Pandas

We use Pandas for handling time series data. Let's start by loading a dataset and converting a column into a DateTime index.

import pandas as pd

# Load dataset
data = pd.read_csv("sales_data.csv")

# Convert 'date' column to datetime format
data['date'] = pd.to_datetime(data['date'])
data.set_index('date', inplace=True)

# Display first few rows
print(data.head())

3. Resampling and Aggregation

Resampling helps in changing the frequency of time series data. For example, converting daily data into monthly data.

# Resampling to monthly data (sum of sales per month)
monthly_data = data.resample('M').sum()
print(monthly_data.head())

Common resampling options:

• 'D' → Daily
• 'W' → Weekly
• 'M' → Monthly
• 'Y' → Yearly

4. Rolling Statistics (Moving Average)

Rolling mean (moving average) smooths fluctuations in time series data.

import matplotlib.pyplot as plt

# Compute 7-day moving average
data['7-day MA'] = data['sales'].rolling(window=7).mean()

# Plot original data and moving average
plt.figure(figsize=(10, 5))
plt.plot(data.index, data['sales'], label="Original Sales")
plt.plot(data.index, data['7-day MA'], label="7-Day Moving Avg", color='red')
plt.legend()
plt.show()

5. Detecting Trends and Seasonality

Using Seasonal Decomposition to analyze trends, seasonality, and residuals.

from statsmodels.tsa.seasonal import seasonal_decompose

# Decompose the time series
decomposition = seasonal_decompose(data['sales'], model='additive')

# Plot components
decomposition.plot()
plt.show()

6. Time Series Forecasting Basics

A basic forecasting technique is Exponential Smoothing:

from statsmodels.tsa.holtwinters import SimpleExpSmoothing

# Apply Simple Exponential Smoothing
model = SimpleExpSmoothing(data['sales']).fit(smoothing_level=0.2, optimized=False)
data['forecast'] = model.fittedvalues

# Plot the forecast
plt.figure(figsize=(10, 5))
plt.plot(data.index, data['sales'], label="Original Sales")
plt.plot(data.index, data['forecast'], label="Forecast", color='red')
plt.legend()
plt.show()

Conclusion

• Time series data is essential for forecasting and trend analysis.
• Pandas provides powerful tools for handling time series data (resampling, rolling mean, etc.).
• Seasonal decomposition helps identify trends and patterns.
• Forecasting techniques like exponential smoothing provide simple predictions.