Tutorial: Visualising High-dimensional Data with R

Website: https://dicook.github.io/tutorial-highd-vis

Workshops for Ukraine

This is for scientists and data science practitioners who regularly work with high-dimensional data and models and are interested in learning how to better visualise them. You will learn about recognising structure in high-dimensional data, including clusters, outliers, non-linear relationships, and how this can be used with methods such as supervised classification, cluster analysis and non-linear dimension reduction.

Background: Participants should have a good working knowledge of R, and some background in multivariate statistical methods and/or data mining techniques.

Presenter: Dianne Cook is Professor of Statistics at Monash University in Melbourne, Australia. She is a world leader in data visualisation, especially the visualisation of high-dimensional data using tours with low-dimensional projections, and projection pursuit. She also works on bridging the gap between exploratory graphics and statistical inference. Di is a Fellow of the American Statistical Association, past editor of the Journal of Computational and Graphical Statistics, and the R Journal, elected Ordinary Member of the R Foundation, and elected member of the International Statistical Institute.

Structure of tutorial

Background: Participants should have a good working knowledge of R, and some background in multivariate statistical methods and/or data mining techniques.

time	topic
15	Introduction: What is high-dimensional data, why visualise and overview of methods
25	Basics of linear projections, and recognising high-d structure
20	Effectively reducing your data dimension, in association with non-linear dimension reduction
30	Understanding clusters in data using visualisation
30	Building better classification models with visual input

Session 1 Slides

Session 2 Slides

Zip file of materials

Getting started

1. You should have a reasonably up to date version of R and R Studio. Install the following packages, and their dependencies.

install.packages(c("readr", "tidyr", "dplyr", "ggplot2", "tourr", "mulgar", "geozoo", "detourr", "palmerpenguins", "GGally", "MASS", "randomForest", "mclust", "crosstalk", "plotly", "viridis", "conflicted"), dependencies=c("Depends", "Imports"))

Ideally, you install this package from GitHub:

remotes::install_github("casperhart/detourr")

1. Download the Zip file of materials to your laptop, and unzip it.

2. Download just the R scripts, slides1.R, slides2.R

3. Open your RStudio be clicking on tutorial.Rproj.

GitHub repo with all materials is https://dicook.github.io/tutorial-highd-vis/.

Copyright

These materials are licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.