Graphics in R
by Chris Wild
This workshop will be of special relevance to those involved in teaching statistics at all levels (e.g., secondary schools and universities), including the general learning of statistical concepts and exploratory data analysis, especially within a Big Data context.
iNZight is a free easy-to-use, standalone, graphical-user-interface-driven system for dynamic statistical graphics and analysis that gives extremely fast access to a powerful suite of techniques, including seamless provision for data from complex survey designs. It runs on a variety of platforms. We will explore some of the unique capabilities of iNZight and its packages.
Additionally, we will look at iNZight Lite, a browser-driven online software that uses a Shiny interface to serve up the same capabilities. Sitting in behind both are a set of R packages whose functions can also be called directly from R. All these modes of working have their advantages and disadvantages. They can not only suit different types of user, but also the same user in different situations. The online system, for example, has the obvious advantage of needing no installation. Less expected is its ease of use for offering sophisticated visualisation and analysis capabilities to remote and independent data repositories whereby a repository data set automatically opens up inside the tool at the click of a link on the repository’s own web page.
Target audience: anyone who is interested in using dynamic statistical graphics and wants to try something new, or to explore the relative strengths of menu and code-driven graphics in a situation where they are delivering the same pictures. Participants who are not already familiar with R will not need to use R if they do not wish to.
About the facilitator/Instructor
Chris Wild is a Professor of Statistics at The University of Auckland. Over the last two decades he has made large contributions to statistical education, including software development and curriculum changes in New Zealand secondary schools. He has given many keynote/plenary addresses and is a past president of the International International Association for Statistical Education (IASE) of the International Statistical Institute (ISI). He is also a Fellow of the American Statistical Association and Fellow of the Royal Society of New Zealand.
|
Getting to Know Grid Graphics
by Paul Murrell
Requirements:
This course will not be appropriate for complete R newbies. It will be assumed that the audience is familiar with R and comfortable writing R expressions. On the other hand, no statistical or graphical expertise will be required.
There will be a series of exercises, so it is recommended that attendees bring their own laptop, with R (and ggplot2) installed.
Outline:
The ‘grid’ graphics package provides a low-level graphics system for R. Many R users do not have direct contact with ‘grid’, but they regularly make indirect use of ‘grid’ whenever they draw a ‘lattice’ or ‘ggplot2’ plot.
This tutorial will expose the ‘grid’ graphics system that is lurking behind higher-level packages, like ‘lattice’ and ‘ggplot2’, and explore the tools that ‘grid’ provides to modify, customise, reuse,
and augment those higher-level plots. We will learn about the fundamental concepts in ‘grid’ graphics—grobs, units, and viewports—and we will learn functions to explore, access, and
manipulate grobs and viewports.
About the facilitator/Instructor
Paul is an Associate Professor in the Department of Statistics at The
University of Auckland. He is a member of R-core, the development
team for the core R system, with a focus on the R graphics system. He
is the author of several R packages, which also have a strong graphical
theme, including ‘gridGraphics’, ‘gridSVG’, and ‘grImport’. He is also
the author of the book “R Graphics”, which describes the core R
graphics system in excruciating detail.
|
Analysing spatial point patterns using spatstat
by Rolf Turner
Prerequisites:
- familiarity with R
- some knowledge of the basics of statistical inference
- an interest in spatial point pattern data
Outline:
This workshop will provide a brief overview of some of the topics covered in the book “Spatial Point Patterns: Methodology and Applications with R” by Adrian Baddeley, Ege Rubak and Rolf
Turner (CRC Press/Chapman & Hall 2016).
I will focus on a few of the fundamental topics in this area of study. These will include:
- the nature of point pattern data
- the importance of the “observation window”
- the nature of marks and covariates
- Poisson processes
- modelling trend
- some interaction models; Gibbs models; clustering models
- summary functions and their envelopes
- inference for spatial point process models
- Monte Carlo tests
I will *briefly mention* some more “cutting edge stuff”:
- replicated point patterns
- hierarchical interaction models
- point patterns on linear networks
About half the time will consist of lecturing and demonstration. The other half will consist in participants’ undertaking computer exercises on the analysis of spatial point patterns.
All participants will need to bring a laptop, with R and spatstat (latest versions!) installed, to the workshop. If you have (one or more) point pattern data sets that you are interested in analysing you are welcome to bring them along on your laptop and I would be happy to discuss the analysis with you.
About the facilitator/Instructor
Rolf Turner is a co-author of the well-received book “Spatial Point Patterns: Methodology and Applications with R” (CRC Press/Chapman & Hall 2016). He is an Honorary Research Fellow in the Department of Statistics at the University of Auckland, in which he taught for several years a graduate course on spatial point processes.
Rolf has considerable expertise in statistical computing and has worked as a statistician in the Division of Mathematics and Statistics at CSIRO, the University of New Brunswick, and the Starpath Project at the University of Auckland. Rolf hails from the Northwest Territories of Canada. He is thus one of the very few mathematical statisticians able to claim to be from “north of 60”.(One other is Robert Gentleman who is of course one of the authors of R. Robert is from Whitehorse in the Yukon.)
|
Faster R code
by Thomas Lumley
This workshop will cover some intermediate and advanced techniques
for optimising R code.
- Profiling vs premature optimisation
- Fast and slow data structures in memory and on disk
- Vectorisation: speed/storage tradeoffs
- Basic parallel processing.
- Better algorithms: stochastic SVD as an example
About the facilitator/Instructor
Thomas Lumley is Professor of Biostatistics at the University of Auckland, and a member of the R Core team. He is also a Fellow of the American Statistical Association and Fellow of the Royal Society of New Zealand. His highly-downloaded R packages include leaps, biglm, survey, and dichromat.
|
