Time-Series Analysis

Summary

This course teaches about time-series analysis and the methods used to predict, process, and recognize sequential data. Topics include:

  • An introduction to time-series and stationary data
  • Applications such as data smoothing, autocorrelation, and AutoRegressive Integrated Moving Average (ARIMA) models
  • Advanced time-series concepts such as Kalman filters and Fourier transformations
  • Deep learning architectures and methods used for time series analysis

By the end of this course, students will have practical knowledge of:

  • Time-series analysis theory and methods
  • Key concepts that include filters, signal transformations, and anomalies
  • How to use deep learning, autocorrelation, and ARIMA with Python*

The course is structured around eight weeks of lectures and exercises. Each week requires three hours to complete. 

Prerequisites

Python programming

Working knowledge of pandas and scikit-learn*

Basic statistics

Week 1

This class introduces time series and its applications. Topics include:

  • What time series is and why it is important
  • How to decompose trend, seasonality, and residuals
  • What additive, multiplicative, and pseudo-additive models are
  • The application of time series forecasting with Python
Week 2

This class introduces stationarity and its mathematical transformations. It includes:

  • The definition of stationarity and its relevance
  • Transformation methods such as differencing, detrending, and logarithms
  • How to differentiate nonstationarity and stationarity data with Python
Week 3

This class teaches about data smoothing methods and their applications. Learn about:

  • Why data smoothing is essential for data analysis
  • Data smoothing techniques—from simple average to triple exponential smoothing
  • How to smooth time series data with Python
Week 4

This class explains autocorrelation and partial autocorrelation. Topics include:

  • What autocorrelation and partial autocorrelation functions are and how they work
  • The variations of models such as autoregressive and moving average models
  • How to use Python to build autocorrelation models
Week 5

This class introduces AutoRegressive Moving Average (ARMA), ARIMA, and Seasonal AutoRegressive Integrated Moving Average (SARIMA) models. Topics include:

  • How ARMA, ARIMA, and SARIMA models work and how to build them
  • How to implement these models with Python
Week 6

This class goes into further detail about advanced time series. Topics include:

  • How to use control charts for anomaly detection
  • An introduction and use case for Kalman filters
Week 7

This class introduces signal transformations. Learn about:

  • Why signal transformations are useful for time series analysis
  • Techniques such as Fourier transformations, filters, and window functions
Week 8

This class teaches how to use deep learning with time series analysis. Topics include:

  • An explanation of recurrent neural network (RNN) and long short term memory (LSTM) architectures
  • How to use Python to implement deep learning models for time-series forecasting