Learning Objectives

Following this assignment students should be able to:

• integrate programing fundamentals and working with data
• solve a data analysis problem with logical and automated code chunks
• communicate effectively with informative and well-styled R scripts

Reading

• Topics

  • Style
  • Debugging

• Readings

  • Hadley Wickham’s R style guide

Lecture Notes

1. Debugging
2. Project Structure

---

Exercises

1. -- Size-biased Extinction --

   There were a relatively large number of extinctions of mammalian species roughly 10,000 years ago. To help understand why these extinctions happened scientists are interested in understanding whether there were differences in the body size of those species that went extinct and those that did not.

   To address this question we can use the largest dataset on mammalian body size in the world, which has data on the mass of recently extinct mammals as well as extant mammals (i.e., those that are still alive today). Take a look at the metadata to understand the structure of the data. One key thing to remember is that species can occur on more than one continent, and if they do then they will occur more than once in this dataset. Also let’s ignore species that went extinct in the very recent past (designated by the word "historical" in the status column).

   Import the data into R. If you’ve looked at a lot of data you’ll realize that this dataset is tab delimited. Use the argument sep = "\t" in read.csv() to properly format the data. There is no header row, so use head = FALSE. The unknown value used in the dataset is -999. R assumes your unknown value is NA, but "NA" in the data is the code for North America. Use the additional arguments stringsAsFactors = FALSE, na.strings = "-999" in read.csv() to get R to keep "NA" as a string and transform -999 to NA.

   It’s probably a good idea to add column names to help identify columns:

   colnames(mammal_sizes) <- c("continent", "status", "order", 
   "family", "genus", "species", "log_mass", "combined_mass", 
   "reference")
   

   1. Create a new RStudio project and a new version control repository for this exercise and commit your changes in small logical chunks. Make sure to commit
      all of the files that are needed for the analysis to the repository. If you would like a private repository on GitHub please ask your instructor to set
      one up. Throughout the assignment focus on using good style to make the code easy for you or someone else to read.
   2. Calculate the mean mass of the extinct species and the mean mass of the extant species. Don’t worry about species that occur more than once. We’ll consider the values on different continents to represent independent data points.
   3. It looks like the species that went extinct are larger on average, but there are lots of different processes that could cause size-biased extinctions so it’s not as informative as we might like. However, if we see the exact same pattern on each of the different continents that might really tell us something. Repeat the analysis but this time compare the mean masses within each of the different continents (dplyr would be one way to do this). Export your results to a csv file where the first entry on each line is the continent, the second entry is the average mass of the extant species on that continent, and the third entry is the average mass of the extinct species on that continent (spread() from tidyr is a handy way to convert the standard dplyr output to this form). Call the file continent_mass_differences.csv.
   4. Looking at the averages was a good start, but we really need to look at the full distributions of masses of the two groups to get the best picture of whether or not there was a major size bias in extinctions during the late Pleistocene. Make a graph that shows the data for each continent that you think is worth visualizing. For each continent display two histograms (these can be on the same axes or separate sets of axes) that use the same bins to display the number of extinct and extant species. Use the log_mass rather than the mass itself so that you can see the form of the distributions more clearly. facet_grid or facet_wrap may be useful to laying out the subplots. Label the plots to make it clear to someone viewing them what they are looking at. Save the graph or graphs as .png file(s) (this should happen automatically in the code).
   [click here for output] [click here for output] [click here for output]