GitHub

Time Series Analysis with R

Course Level: Intermediate

Predicting the future is a tough problem. Time series analysis makes it possible to assess whether or not predictions are possible and, if they are, build a model which can generate informed predictions for the future with realistic estimates of uncertainty. This training course will introduce participants to the packages in the Tidyverts.

The best qualification of a prophet is to have a good memory – George Savile

Book: Time Series Analysis with R

 Start Date: Price: Venue Details: Time: Duration:

Course Details

• Course Outline
• Learning Outcomes
• Materials
• Prior Knowledge
• Introduction to tsibbles: Using the {tsibble} package to manipulate time series data
• Features and Visualisation: Creating seasonal, lag and autocorrelation plots using the {feasts} package
• STL Decomposition: De-constructing a time series into it’s seasonal and trend components
• Introduction to forecasting: Constructing simple forecasts with the {fable} package
• Exponential Smoothing: Creating and forecasting with ETS models
• ARIMA models: Creating ARIMA models and forecasting

Session 1:

By the end of session 1 participants will…

• have an understanding of what time series are and be able to store time series data in R using {tsibble}.
• know how to visualise time series data using {feast} and {ggplot2} for seasonal plots, subseries plots, lag plots and autocorrelation.
• gain knowledge of time series decomposition and be able to fit and models using STL decomposition and plot the components. - have the ability to choose STL model parameters and be able to acquire seasonally adjusted series.

Session 2:

By the end of session 2 participants will…

• be familiar with different forecasting methods.
• be able to use the {fable} to create forecasts and {feasts} visualise them.
• know how to extract prediction intervals from a forecast, get the residual forecast plots and determine accuracy of a forecast.
• have an understanding of exponential smoothing for modelling time series.
• be able to apply SES models.
• have an understanding of ARIMA models and have the ability to fit, forecast and visualise both non-seasonal and seasonal ARIMA models.

This course does not include:

• an in-depth discussion of the statistical principles behind the forecasting methods covered.
• regression-based models - see our Tidymodels course instead.
• spectral methods.

This course assumes basic familiarity with R and the {tidyverse}. Attending our Getting to Grips with the Tidyverse course, is more than sufficient in providing you with the prerequisite knowledge required for this course!

Attendee Feedback

• “Very beneficial course helped by insightful and clear explanations from the trainer.”
• “This is my second course with Rhian and she is a great and clear instructor. She is engaging, even virtually, and I find the structure of the course very helpful for learning. Short lecture, followed by a demo, followed by a practical and then coming together to discuss is a great way to make the material stick. Thank you Rhian!”