Course summary

The “Time Series Analysis and Forecasting with R” tutor-led online training course will provide you with essential knowledge to allow wrangling, processing, analysis and forecasting of time series data using specialised libraries such as ts, xts, zoo, tsibble, prophet, fable and forecast for R programming language. Whether you wish to analyse financial data, predict sales or marketing revenue, or understand temporal patterns in your social, medical or economic data, this course will provide you with theoretical and practical understanding on how to clean, visualise and model time series data in your workflows using R programming language.

During the course, you will first learn to manipulate the imported data, extract necessary date/time stamps and transform the processed data into supported time series R objects. You will then proceed to perform essential time series exploratory and decomposition operations, calculate selected moving/rolling single-value statistics, convert between differing time frequencies, visualise and prepare data for predictions. The forecasting part will include sessions on estimating linear, non-linear and locally-weighted trends, multiple regression models, ARMA and ARIMA approaches, dynamic models and a selection of machine learning and AI methods applicable to time series data e.g. Support Vector Machines and Long-Short Term Memory deep learning methods.

A mixture of online pre-recorded instruction videos, weekly live group webinars with our tutor, additional 1-2-1 check-ins (either via email or on Microsoft Teams) and several homework exercises throughout the duration of the course will ensure you will be able to apply Machine Learning methods using R language to your own data and research questions.

Who is this course for?

This tutor-led online course is suitable to all data scientists, researchers, Master’s or PhD students, post-doctoral researchers, data analysts, developers and data engineers, who are currently using R language (preferably at intermediate level) and would like to expand their skills to include industry-standard and more cutting-edge time series forecasting methods.

What is included in the course fee?

At Mind Project, we pride ourselves for delivering great quality, goal-directed, interactive and cost-effective training. Our remote, hybrid, and on-premises tutor-led courses are designed to maximise your opportunity to learn complex topics in shortest possible time while being able to discuss and interact with our tutor and other fellow attendees. Therefore, the registration fee for this course includes a number of benefits that many other training providers don’t offer:

• Weekly (90-minute long) remotely delivered live group webinars with our expert tutor over the period of 4 weeks – you will be able to ask questions, interact and discuss the topic with other attendees from the comfort of your home/office,
• you will have access to course online materials for 1 year via our Mind Project Learning Platform e.g. tutorial videos, exercises and quizzes, R code scripts used during tutorials and live sessions, example datasets, optional and mandatory reading (e.g. blog articles, academic papers, industry reports), and other external recommended resources e.g. online books,
• 2-week post-course follow-up period during which you will be able to work on homework exercises and submit them for our feedback – the feedback will be provided either via email or Microsoft Teams (i.e. as a 15-minute 1-2-1 session with our tutor),
• exercises and tasks completed by you between the live webinar sessions will help you in better material retention and will enhance your learning progress,
• email supervision and weekly check-ins during the course period will maximise your learning outcomes and improve your learning experience; we will provide you with support throughout the course and the 2-week post-course follow-up period – feel free to ask us questions about the material presented at the training during this period,
• small group size allows easy interaction and stimulating environment for successful learning.

Programme outline

During this course, you will learn and implement a variety of data analysis and forecasting methods to explore and predict the patterns in time series data including linear trend, seasonal and exponential smoothing models, non-seasonal and seasonal ARIMAs, dynamic regression models as well as more advanced predictive analytics methods such as Support Vector Machines and Long Short-Term Memory approaches for multivariate time series forecasting. The course will be run according to the following schedule:

Week 1: Working with time series data in R

• Challenges with time series data with R,
• Creating time series objects and data structures,
• Converting between different time series objects,
• Importing time series data,
• Pre-processing time series data,
• Converting between date/time types,
• Extracting specific components of data and time,
• Plotting time series data with ggplot2 and interactive visualisation packages e.g. plotly and highcharter,
• Downsampling and upsampling time series,
• Handling time series missing values,
• Exploratory analysis of time series data.

Week 2: Time series analysis with R

• Building on exploratory analysis of time series: moving averages, lagged values and rolling statistics,
• Time series decomposition methods,
• Autocorrelation, stationarity and differencing,
• Transformations and adjustments,
• Using decomposition for forecasting,
• Evaluating forecasting accuracy,
• Simple forecasting approaches: naive model, average model, linear trend model.

Week 3: Univariate time series forecasting methods

• Introduction to univariate time series methods: simple exponential smoothing, Holt’s linear trend and Holt-Winter’s seasonal methods,
• Exponential smoothing state space models,
• Autoregressive (AR) and moving average (MA) models, non-seasonal and seasonal ARIMAs,
• Facebook’s prophet library for univariate time series forecasting,
• Introduction to deep learning for univariate time series with Long Short-Term Memory (LSTM).

Week 4: Multivariate time series forecasting methods (and AI)

• Multiple linear regression with time series data,
• Polynomial regressions with time series data,
• Combining ARIMAs with multiple linear regressions: dynamic regression models,
• Support vector machines (SVMs) and kernel smoothing methods with multivariate time series,
• Long Short-Term Memory for multivariate time series forecasting.