Papers and Talks


Talk: A Shifting Landscape for Laboratory Science,   R Markdown source   (Mt Albert Research Centre, May 10 2016)
The slides comment on issues with published content more generally --- Reinhart-Rogoff: "Growth in a Time of Debt"; Jung et al: "Female hurricanes are deadlier than male hurricanes".
README file (has details of other files in the directory, including R Markdown source files)
Talk: Reproducibility in Science --- Rethink, or Crisis,   Shorter version (has some new content)
(Victoria University of Wellington, Wed 4 Nov 2015)
Statistics for a More Scientific Science. How much of published laboratory science should one believe?
  Substantially expanded version of talk given at NZSA+ORSNZ 2014.
  NB: Supplemental materials may relate to an earlier version of the talk.   See above for a later version. Supplementary materials.
Technology-enabled advance in the worlds of statistics, machine learning and data mining.
  New Zealand Science Review 68 (2011): 113-116.   Supplementary materials.
Interview with Steve Miller of Business Intelligence Network Newsletter
Data, science and new computing technology.
  New Zealand Journal of Science 62 (2005): 126-128.
The Role of Models in Predictive Validation (Statistics for Budding Data Miners.
  (Invited paper, Session 69, ISI meeting, Berlin 2003     Overheads)
Florence Nightigale on: 1) The importance of statistics; 2) Social science as theology.
The Use of Generalized Additive Models in Spatial Interpolation

Data Mining coursenotes, talks and papers

Lecture Notes, etc.:

Least squares computation; Theory of Generalized Linear Models; Regression in Practice;
Comparative a/c of lda(), qda() and logistic regression; Spatial methods in R; Various other notes.

List of Selected Publications: 1984-2007 (pdf file)

Notes on the R System

Using R for Data Analysis and Graphics -- Introduction, Examples and Commentary
(pdf file, 96 pages, minimally revised Jan 2008)
Lattice graphics: use of functions (from latticeExtra) that build up graphs in layers; Control of bounding boxes.
An R-based Interface to the Google Visualisation API (use, e.g., to create Hans Rosling style Motion Charts).
R-based Interfaces to Google Maps

Research Planning & Methodology

The Design of Research Studies - A Statistical Perspective
The Rationale of Research (18pp.)
Notes on Experimental Design

Older Materials