Statistical Computing 1

R is a high-level programming language widely used for statistical computing. Base R provides users with a large variety of statistical and graphical tools, which are easily extended via the use of packages. R is an interpreted programming language. This means that R users execute code via a command-line interpreter and code is interpreted into low-level instructions one line at a time. Integrated development environments (IDEs) such as RStudio provide easy access to such an interpreter through the use of a console....

When producing research which involves a lot of code, it is important that your research/analyses can be reproduced. When analyzing data it is common to write some code, change it a little and test some version of your code in the console. Then when your final document is produced, the R script or Markdown document may not produce the exact same results as your published analysis. This is one of the main challenges of producing reproducible research....

R makes it relatively easy to structure a project, by creating a “project”. It also integrates git and github which makes it easy to track changes and publish code to the web. Projects: organizing code The best way to approach projects is to maintain a well-defined file structure. RStudio projects allow you to easily divide your work into different directories. Of course, this can be done manually. Generally a good high-level file structure within your project’s directory is:...

Vectorization Vectorization is the process of replacing the use of loops in code, with a single operation applied elementwise to the vector. In vectorizing code, the multiple instructions running the same operation in a for loop are replaced by a single instruction which applies the operation to multiple elements in a vector. Vectorization is much faster than using for loops in interpreted languages such as R, Matlab, and Python. To see this in practice, we will look at some examples....

Programming paradigms R supports multiple programming paradigms. Programming paradigms describe the way in which code is written and structured for larger projects. The types of programming structures are: Imperative programming — We define a sequence of instructions that modify the ‘state’ of the program. Object-oriented programming — Objects combine the machine’s state with a set of methods. Programs are then defined via the construction of objects. Declarative programming — A sequence of operations is defined that only specify what the output of the program should be....