NHS-R Workshop: Forecasting Using R – Advanced Methods – May 2023

Forecasting Using R – Advanced Methods – 1 day online workshop

• Wednesday 24th May 2023

The course duration will be one day from 09:00am am to 3:00 pm and will take place online.

Course description

It is becoming increasingly common for organizations to collect huge amounts of data over time with finer time granularity such as arrival time (date-time), and existing time series analysis tools are not always suitable to handle the scale, frequency and structure of such collected data. In this 1 day workshop, we will look at some advanced methods that have been developed to handle forecasting of large collections of high frequency time series such as sub-daily, build models to include time series patterns and relevant predictors and introduce methods to forecast hierarchical and grouped time series structures.

Learning objectives

Attendees will learn:

1. How to forecast time series data with long and multiple seasonal cycles.
2. How to include subtle time series dynamics and predictors using dynamic regression models
3. How to forecast a large collection of time series with a hierarchical/grouped structure.

Is this course for me?

This course will be appropriate for you if you answer yes to these questions:

1. Do you know how to use functions in R?
2. Do you already use R for forecasting?
3. Do you need to forecast large collections of related time series?
4. Do you deal with time series with long and multiple seasonality such as weekly, daily and sub-daily?
5. Do you want to build models to include time series data and relevant predictors?

Prework

People who don’t use R regularly for forecasting, are recommended to have a look at chapter 1-8 of the fpp3 book at https://otexts.com/fpp3/ beforehand.

Please bring your own laptop with a recent version of R and RStudio installed. The following code will install the main packages needed for the workshop.

install.packages(c(“tidyverse”,”fpp3″, “lubridate”, “GGally”, “sugrrants”, “astsa”))

Schedule

Slides

Session 1: Forecasting time series with long and multiple seasonal cycles

Session 2: Dynamic regression

Session 3: Forecasting a collection of time series using hierarchical and grouped structures

Please get in touch if you have any questions, email: nhs.rcommunity@nhs.net

Dr. Bahman Rostami-Tabar

Associate Professor in Data & Management Science

Cardiff Business School, Cardiff University

Chair and founder, Forecasting for Social Good, International Institute of Forecasters

www.bahmanrt.com

Registration

Please register via the Sign up link in the right panel.
You will need to register as a member of the NHS-R Community to register for NHS-R Events.

You will receive a confirmation email, further workshop information including the calendar invite with the Zoom link to join the workshop will be sent out to you nearer the date.

Please note: This course is limited to NHS employees and those who work in UK public sector organisations such as PHE, local councils and DHSC.

For further information, please contact: nhs.rcommunity@nhs.net