Contextual Task Analysis

This contextual task analysis is based on what is believed most users do with the current R interfaces, but in an abstract way (could apply to many other "statistical analysis" activities). We then alter the work model developed from this analysis to include obveRsive and incorporate what it does to the users activities. In this way we are able to analyze whether obveRsive does give the user an advantage and whether it fits into the context we intend.

Current R Usage

The current way of using R is generally broken down into three basic patterns:

1. Start R "command line interface" and type various commands to it while interactively analyzing the data and outputs.
2. Using an external editor, write R scripts inside full files in a way similar to other programming languages and run them through the interpreter in a non-interactive fashion.
3. A hybrid of the two where you write R scripts to support activities with the command line interface.

ObveRsive is really only interested in helping people who deal mostly in realm #1. While it may help with those working in #2 and #3, it is primarily targetted at people who use R in an interactive fashion.

Contextual Analysis

After working with R for various data analysis activties, we have concluded that the interactive operations with R follow this basic abstract model:

From this basic task analysis, you can see that a major problem comes from the user being required to do everything themselves, with very little help from R. Each of the primary actions in response to each event is carried out by the user specifically and usually involves a fair amount of recall and work to complete. This "manual intervention" required by the user for such mundane tasks as data viewing and transformation is where the most help is needed. The analysis of data will always be very hands on and not easily automated, but much of the tedium can be eliminated through "wizard" interfaces and other tools.

ObveRsive can make the greatest contribution by either doing much of this work for the user, or simply reducing the number of things a user needs to remember or do for each action. Take data viewing for instance. In R it is necessary to remember various function calls, the proper orders of parameters to them, to save the results of these functions to new variables, etc. By providing a simple "spreadsheet" interface the user simply has to view or edit the data. They are not required to remember very much and can use an interface that is already familiar to them.

Proposed ObveRsive Usage

As mentioned previously, ObveRsive can best help the user by relieve as much of the "memory recall pressure" as possible through familiar interfaces. If obversive is able to take on the majority of the work needed to View and Transform Data, Create and Export Plots, and provide higher level interfaces for Data Anlysis, then the activities shown in the above diagram will become easier.