Time series analysis patterns in Pandas

Introduction

Time series analysis helps us understand data that changes over time. It shows patterns like trends and cycles to make better decisions.

Tracking daily sales to see if business is growing

Monitoring temperature changes over months

Checking website visitors hour by hour

Analyzing stock prices to find trends

Observing heart rate data during exercise

Syntax

Pandas

import pandas as pd

# Load time series data with a datetime index
series = pd.Series(data, index=pd.to_datetime(dates))

# Common pattern checks
series.plot()  # Visualize data
series.rolling(window=3).mean()  # Moving average to smooth data
series.diff()  # Differences to find changes
series.autocorr(lag=1)  # Correlation with lagged data

Time series data must have a datetime index for proper analysis.

Rolling windows help smooth short-term fluctuations to see trends.

Examples

This example creates a simple daily time series and plots it to see the pattern.

Pandas

import pandas as pd
import matplotlib.pyplot as plt

dates = pd.date_range('2024-01-01', periods=6)
data = [10, 12, 15, 14, 13, 16]
series = pd.Series(data, index=dates)

# Plot the time series
title = 'Daily values'
series.plot(title=title)
plt.show()

Calculates the 3-day moving average to smooth the data and highlight trends.

Pandas

rolling_avg = series.rolling(window=3).mean()
print(rolling_avg)

Finds the difference between each day and the previous day to see changes.

Pandas

diffs = series.diff()
print(diffs)

Measures how much today's value relates to yesterday's value.

Pandas

autocorr_value = series.autocorr(lag=1)
print(f'Autocorrelation with lag 1: {autocorr_value}')

Sample Program

This program creates a time series, plots it, calculates a moving average to smooth the data, finds daily changes, and measures autocorrelation to see how values relate over time.

Pandas

import pandas as pd
import matplotlib.pyplot as plt

# Create sample time series data
dates = pd.date_range('2024-01-01', periods=10)
data = [5, 7, 6, 8, 10, 9, 11, 13, 12, 14]
series = pd.Series(data, index=dates)

# Plot original data
series.plot(title='Original Time Series')
plt.show()

# Calculate 3-day moving average
moving_avg = series.rolling(window=3).mean()
print('3-day Moving Average:')
print(moving_avg)

# Calculate daily differences
diffs = series.diff()
print('\nDaily Differences:')
print(diffs)

# Calculate autocorrelation with lag 1
autocorr = series.autocorr(lag=1)
print(f'\nAutocorrelation (lag=1): {autocorr:.2f}')

Important Notes

Missing values can appear in rolling averages at the start because there is not enough data to fill the window.

Autocorrelation helps detect repeating patterns or trends in time series.

Plotting is a great first step to visually understand your time series data.

Summary

Time series patterns show how data changes over time.

Use rolling averages to smooth data and see trends.

Differences and autocorrelation reveal changes and relationships in time.