Forecasting using R- NHS

  • Prepare data for forecasting
  • Determine what to forecast
  • Identify systematic patterns in time series
  • Apply Exponential Smoothing, ARIMA and Regression models
  • Produce point forecasts and prediction intervals
  • Evaluate the accuracy of forecasting models
  • Visualize, export and report result for interpretation and insights

Forecasting in R • October 2020 • Bahman Rostami-Tabar • Cardiff University


20 April 2021

10:00 - 16:00

27 April 2021

10:00 - 16:00

04 May 2021

10:00 - 16:00



This is a tentative course schedule. The flow of topics might change slightly depending on how quickly / slowly it feels right to …

Pre session- Prepare data for forecasting using tsibble

The tsibble package provides a data infrastructure for tidy temporal data with wrangling tools. Given that the fable package will be …

Session 1

Learn about key features of time series and how to use time series graphics to identify them.

Session 2

We learn some key tools in forecasting that are useful for many forecasting situation. these tools will be used in any forecasting …

Session 3

You will learn how to use expoentnial smoothing to capture both trend and seasonality available in time series and use it to forecast …

Session 4

ARIMA models provide another approach to time series forecasting. While exponential smoothing models are based on a description of the …

Session 5

The basic concept is that we forecast the time series of interest assuming that it has a linear relationship with other time series.


Course components


Pre Session Data preparation in tsibble Reading and video
Session 01-1 Introduction to forecasting Tuesday 20 April 2021
Session 01-2 Time series patterns and graphics Tuesday 20 April 2021
Session 01-3 Forecaster’s toolbox Tuesday 20 April 2021
Session 02 Forecaster’s toolbox: modeling accuracy evaluation Tuesday 27 April 2021
Session 03 Exponential smoothing models Tuesday 27 April 2021
Session 04 ARIMA models Tuesday 04 May 2021
Session 05 Regression Tuesday 04 May 2021

Online sessions

Given that all sessions are online, they need to be as interactive as possible. My role as instructor is to introduce you new tools and techniques, but it is up to you to take them and make use of them. It is best if you can start applying what you learn in the workshop on your own data.


You will also complete tutorials in R for each session.

You need to use RStudio Cloud for all lab sessions.

Please click on this link to connect to RStudio Cloud workspace.

We will also share a link with you on the day through the Slack.

Sharing / resing materilas

Feel free to reuse and share materials. If you use any material from ,please endorse the source.


We will be using Slack to facilitate course communication, particularly around questions and answers. If you have a question outside of class, feel free to reach out via email.


Most of you will need help at some point. Please send an email to Dr. Bahman Rostami-Tabar if you need any help after the workshop or if you have any questions about MSc dissertation, PhD or upcoming webinars.

MSc dissertation

We - Cardiff University and University of Lancaster, are also offering free resources for 3 months through MSc. dissertation projects for the following topics: data science/analytics/ forecasting and their links to decision making. If you would like help with any project, please send me an email.


If you would like to pursue a PhD in healthcare related to management, planning and decision making, logistics, forecasting, analytics or data science, please get in touch.


We start a series of webinar on specific topics from November 2020:

  • Forecasting high frequency time series with multiple seasonalities
  • Forecasting for multiple level of granularity
  • Modern regression in healthcare
  • Application of ordinal regression on healthcare forecasting

If you have any suggstion , please get in touch.

  • 3 colum Drive, Cardiff Business School, CF10 3eu
  • Email


Main instructor


Bahman Rostami-Tabar

Main Instructor



Ivan Svetunkov