Why R for Data Analysis?
All, as in 100%, of data analysis in the modern world is done on the computer. We use R as the app for data analysis, the same app used by many if not most data scientists. R is emerging as one of the primary platforms for data analytics in the world, if not the primary platform. Besides providing a vast array of statistical analyses, R is free as in zero cost, open source, and cross-platform [1:34 video introduction].
R vs. Excel
Both Excel and R analyze data from function calls. Each function accomplishes a specific analysis, such as computing the arithmetic average. To do so, use the Excel function average, or the R function called mean. The distinction is that with Excel the function call is entered into a cell in the worksheet, in this example the 8th cell in Column E. Specify the data to be analyzed by the function with a cell range, such as the relative cell range E2:E7. Relative to the cell in which the function call is entered, this range refers to the cells in the same column that are from six to one cells above it.
In R enter the function call at the command line, in response to the > prompt. Specify the data to be analyzed by a variable name. Here generate the histogram of Salary. Note that the variable name Salary is in the first row of the corresponding Excel worksheet, with the data values listed under the variable name, beginning in the second row.
Following are reasons why professional data analysts use R for their analyses instead of Excel.
- As a spreadsheet app, Excel is great for data entry, but provides only a smattering of very basic statistics computations
- R does Big Data, easily handling data sets with millions of observations, limited only by the computer's available memory
- Once the concept of working with an app like R is understood, less time is required to conduct an analysis, even with the data already in an Excel worksheet, entering a simple command such as Histogram(Salary) instead of clicking and mousing around
- Obtain each R analysis in this course with a single, brief line of code, a call to a specific function, instead of irrecoverable mouse clicks, so the results of R analyses are reproducible
Reproducibility
The list of R function calls to perform an analysis, the lines of code, document exactly how you conducted the analysis. Although our work in this class is simple enough that you do not need to save your code in a separate file (though still to be encouraged), in the real world of data analysis no one enters code directly into R. Instead the R function calls are written into a file to maximize reproducibility, and then run in R whenever desired, by yourself or by others. The instructions for analyses done by one person become accessible to all applicable members of your organization at any subsequent point in time, including yourself, instead of disappearing into digital dust as do mouse clicks.
lessR Enhancements
The downside of using standard R is that it is geek oriented. To make things work, you really do have to enter the exact syntax, and many lines of code are typically required to perform any analysis. One example is that R is case sensitive, which means that capitalization must be followed exactly.
I have made using R for basic data analysis much easier with my 43 or so functions that can be added to the standard R functions. These functions, along with much more extensive and helpful error diagnostics that my functions provide, result in a more or less "un-geeked" R. The function calls are straightforward, but you should always have a means to access them, such as my online help system with the lessR function Help().
My functions are organized into what the R system refers to as a package, which I call lessR. The R system, which is available on servers around the world, consists of the hundreds of base R functions included with the installation of R, plus the functions found in additional packages accepted into this R system. My lessR package is part of this larger R system. We will mostly use my lessR functions in this course, such as the Histogram function, but will be working in the standard R environment.
RStudio
Many (most?) data analysts prefer to run R from within another application called RStudio, an environment for running R. Note that for this course there is no need to run R within RStudio as you can just run the R app directly. RStudio, however, provides additional features and is also freely available: RStudio Desktop. One advantage, for example, is that RStudio makes it easy to transfer R instructions stored in a separate file to run in R, illustrated in my quick introduction to RStudio.