## Recent presentations

Priors for t and skew-t models, with a focus on flexible latent variable models. The source code is also available.

Multivariate Data Analysis via Stan Modelling Software, with a focus on flexible latent variable models. The source code is also available.

Presentation for StatCon2020: Stan as an all-purpose modelling solution (video version with my voice). Also available is the slides and source code zipped up.

## Important Side Topics

Covid-19 Waves by Province in South Africa

## Recent publications

## Older (but still relevant) presentations

Departmental presentation: Three mildly interesting consultation projects

Discussion with computer science honours students

Presentation on the magic of markdown for SASA 2019, including the example of doing an assignment where each student gets unique data and their own memo and the example of incorporating interesting models with STAN into the research or consultation workflow. Source for the examples are available here and here respectively, and data here, while the source for the presentation itself can be downloaded here.

Department presentation on statistical modelling with Stan

Flash facts presentation for UFS discussion

Department presentation on Adventures in Statistical Consulation

## Random topics

Distribution visualisation app, useful for undergrad stats.

Useful code for teaching and research.

My EXCEL addin with tons of useful functions like copying tables in R or LateX format, simulating many distributions, logit transforms, deleting alternating columns, *etc.* Must be manually placed in C:\Users\You\AppData\Roaming\Microsoft\Addins and then activated in EXCEL via File->Options->Addins->Go.

Self administered module evaluations

List of Sean’s Shiny apps *warning: these may be under construction at any time*

Prof. van Zyl asked me to check the formula for the asymptotic variance of the sample kurtosis on Wikipedia empirically. This is an excellent introduction to the idea of a simulation study. It is also a good example of *R markdown* if you look at the source here. Connected to this is a deeper study of sample kurtosis for Stable distributions, in which we show that $E\left[\frac{\text{excess sample kurtosis}}{\text{sample size}}\right]\approx 1-\frac{\text{tail index}}{2}$.

This study has now been published open access and is available at the South African Statistical Journal website.

Here’s an interesting basic probability problem solved with R Markdown.

## Older research work

Publications and reports related to my PhD work

### Older Publications

- An empirical study to find an approximate ranking of citation statistics over subject fields
- Student success: data mining measures what matters

### Reports (not peer reviewed)

- Time Series Analysis of the Southern Oscillation Index using Bayesian Additive Regression Trees
- Variations on Goodness-of-Fit Tests for the Generalized Pareto Distribution
- Bayesian Extreme Value Analysis of Stock Exchange Data
- Determinants of Private Investment in South Africa: Investor Confidence Index Approach
- On Determining the Distribution of a Goodness-of-Fit Test Statistic [Previous title: A Method for Testing Models with Unknown Parameters]
- Selecting an optimum threshold with the Kullback-Leibler deviance measure