Geoff Robinson and Jeromy Anglim have kindly agreed to share their experiences with us on how they go about implementing reproducible research, and what workflows they follow. Thanks guys!
I think this will be a fun, exciting and interactive meetup - great way to end the year!
If you'd like to join dinner afterwards, there's a separate RSVP for that.
Here's the text from Jeromy's original idea:
A presenter could walk through a complete instance of reproducible research.
The content could be a journal article, an industry report, etc.
This could involve Sweave or another approach.
Useful things to show:
- the R code editing environment
- the organisation of files and code
- talk through R code for (data input, data transformation, analyses, and output generation)
- talk through decisions and learning process (e.g., how were specific functions and packages discovered, quality control, etc.)
- reproducible integration of output in the final report (e.g., LaTeX, Open Office, Word, etc.)
Geoff adds the following:
What my R code looks and feels like.
I will present 6 examples of R code to illustrate some issues associated with writing functions, organising the R code associated with a report, using R code to prototype computational algorithms, testing code, writing a set of reports tailored for a set of customers, and interfacing R with C or Fortran. My primary concern is to discuss programming style, especially to learn how other people would have proceeded differently.
Some info about Geoff and Jeromy:
Geoff Robinson has a MSc from Monash (1972) and a PhD from University college London (1975), both in Mathematical Statistics. He has worked at the Australian Road Research Board, the Victorian Agriculture Department and CSIRO on a wide range of applied projects, including many concerned with mineral sampling, mineral blending and scheduling of bulk materials handling yards. His use of R is largely for statistical data analysis and as an interface to C code for modelling mineral stockpiles.
Jeromy Anglim is a lecturer and statistical consultant in Psychological Sciences, The University of Melbourne. He's been using R since 2007 and regularly blogs about statistics, R, and reproducible research at http://jeromyanglim.b... (http://jeromyanglim.blogspot.com)