Back in June I wrote about A Complete Example of Reproducible Research with Org. One of the authors of that paper was Arnaud Legrand who has an excellent webinar on reproducible research. Sadly, the video of the talk is in Flash, which makes it unwatchable for many of us. Happily they also have an MPEG4 version but it took a long time to download. Nonetheless, it was worth the wait. It’s a really great talk on why you should be doing reproducible research and a demonstration of some of the tools you can use to help you organize and produce it.
The first half of the talk considers the problems that ordinary research papers can have and how reproducible research can help solve them. Legrand also talks about some problems with implementing reproducible research. In the second half of the talk, Legrand demonstrates a few of the tools that a researcher can use to help with producing reproducible research.
One of these, reprozip, helps you capture the environment that you used to run your software. He also looks at Rstudio and Ipython and shows how they can help produce papers while embedding code and results directly into the final result. His favorite tool, though, is Emacs/Org Mode, which he says is more powerful and versatile. He describes how he and his students keep journals (in Org mode) that record everything they did and the results they got. These journals are invaluable when it comes time to write the paper. If you read Legrand’s complete example paper, you saw an example of this is in action.
The talk is just over 2 hours 45 minutes so you’re going to have to block out time (or several times) to watch. It’s worth it though, so if you have any interest at all in the subject matter, I urge you to take the time. Don’t worry too much about reading everything on his slides because they are also available.
Finally, I’m happy to report that this is the first in a series of Webinars on reproducible research. There a three more currently available and another is planned for some time this month. The others webinars discuss
- Controlling your environment
- Numerical reproducibility
- Logging and backing up your work
If you follow the link to the other videos, you will find a list of topics they plan to cover in the future.