Week 3: Rstudio, Jupyter notebooks, and publication-quality graphics

Background materials:

• Rstudio website
• Rstudio tutorial from UCLA
• Jupyter project
• Jupyter quickstart guide
• Jupyter notebook tutorial

Lecture notes

For the lab this week, we will work through this set of exercises.

If you know ggplot well already and/or have time, work through this tutorial about dplyr.

Useful free resources:

• ggplot tutorial from 2014.

Books:

• R for data science
• ggplot book
• Basically anything by Hadley Wickham is the future of nice stuff for R.

Getting image files ready for publication.

There are a lot of ways to do this! Here’s one that works well.

Background

• R does not make “true” pdf files!! Rather, it generates a wrapper around some other form of graphic file. In our experience, pdf generated from R raise red flags at journals like PLoS.

So, here is what we do:

1. Save graphics as 600dpi TIFF:

tiff('plot.tiff',height=x,width=y,res=600)

This makes a really big file!

2. Use ImageMagick to convert to smaller tiff:

convert plot.tiff -set colorspace RGB -layers flatten -alpha off \
-compress lzw -depth 8  -density 600 -adaptive-resize 4500x2400  plot.compressed.tif

3. Use ImageMagick again to make a final pdf:

convert plot.compressed.tif plot.compressed.pdf

We have a shell script online to automate the process. Get it here.