Don't Panic — R Programming

R is a language and environment for asking questions of data, then leaving enough evidence that somebody else can ask the same questions again. Before this arrangement, an analysis could be a stack of hand calculations, a spreadsheet with mysterious cells, or a console session that vanished when its window did. R gives the calculation a home. It does not, alas, make the calculation correct by wearing a lab coat.

The main trick is objects. A vector holds a sequence of values. A data frame holds a table. A function gives a repeated operation a name. Each expression can turn one object into another, so an analysis becomes a trail: data arrives, values are transformed, a model or plot appears, and the intermediate steps remain available for inspection. This is much more reassuring than a chart that has arrived without its luggage.

The slightly surprising part is that an interactive session is helpful and dangerous in the same way. The console lets you inspect a data frame, test an index, and learn what a function does. It also remembers objects that your final script may never create. Put the import, transformation, model, and output in a script, then run it again from a clean session. That is where a result starts behaving like evidence rather than a lucky photograph.

Packages extend the environment with functions, data, and documentation. CRAN, the Comprehensive R Archive Network, distributes many of them, and Task Views group them by subject. Start with base R's vectors, data frames, indexing, functions, and help. Add a package because it solves a defined problem, not because the package shelf is impressively tall. A package is an extension, not a replacement for knowing what shape your data has.

Read the Intro when you want the map of R's objects and tradeoffs. Use Slides when you want the data-to-result loop in compact form. Keep the Cheatsheet nearby for names and decision rules. Then use the Practice tab to run a complete script that prints a summary and writes a plot. The Quiz can check whether the vocabulary has settled in. From there, the Reference tab points from the official manuals to the package ecosystems that fit your work.

Don't Panic — R Programming

Relevant careers