How deep is reproducibility?

• very deep
• but not too deep to get started

Good, bad and ugly about the web

A one-minute introduction to knitr

• remember ```{r}, ``` and
• `r ` and you are all set
• (optionally) math
• video at https://yihui.org/knitr

A few selected features with demos

• document formats: LaTeX, HTML, Markdown
• graphics (multiple plots per chunk, tikz graphics, animations)
• cache
• all languages in one document

A book example

• http://theanalysisofdata.com

Further examples on the web

• http://rpubs.com/
• http://tengfei.github.com/ggbio
• http://public.opencpu.org/apps/knitr/
• http://www.ats.ucla.edu/stat/R/
• http://vis.supstat.com/

Why blogging is great for science

• http://techcrunch.com/2012/10/01/nyt-election-oracle-fivethirtyeight-on-why-blogging-is-great-for-science/

1. Statistics Are History Foretold
2. Be Honest About Your Bias
3. Bayesian Nuance
4. No Shame In Changing

• http://andrewgelman.com/2012/09/speaking-frankly/

Conclusions

• technology is becoming easier and cheaper (how much is a domain name?)
• there isn’t much to learn (as long as you know R, reproducibility can be close to you)
  • install.packages('knitr')
• record your analysis by code instead of memory

Questions?

• Feedback (Github issues): https://github.com/yihui
• Email: echo oybuwycxy.trpb | tr bopcrbtucwxyc+ exmiaen@iyuih
• Images credit: http://researchinprogress.tumblr.com