Students: Matthew Wareham
Faculty Advisor or Community Project Lead: Kenrick Mock
The goal of the project was to create a plot viewing application that allows dynamic exploration of a data set or parameter space. Mathematica was the language chosen for implementation initially, and this decision turned out to have pluses and minuses. Decent progress was made over the course of the semester, but eventually development was severely hindered, mainly due to the lack of a thorough understanding of the language’s inner workings and best practices. The project will not cease here and is a continued work in progress, with the next step being to become more familiar with Mathematica to decide if it really is the best choice for fulfilling the objectives, and then to reassess current work and move on from there.
In general, plotting a data set is usually a very important first step toward gaining insight about the data. For the purposes of rigorous data exploration, most plotting applications are either too rigid or specific, like the charts of Google/Yahoo Finance, or too static, like the charts of Microsoft Excel. It appears that most general data mining tools focus on statistical algorithms and less on the visualization Of the data. With visualization oriented tools, like curios-IT, many seem to provide a “do it for me” scheme, where the program finds an appropriate plot type that matches the data on its own. The results of these programs are often unfortunately not very dynamic—they are fine for certain purposes, especially if you just need something quick. This project’s goal is to improve upon the data viewing solutions that are currently available, which have a number of issues when it comes to trying to “swiftly” and efficiently explore both the big picture and the intricate detail of a data set. Since 2D and 3D Cartesian coordinate plots are the most widely applicable visualization technique, they will be the first focus, with other types of plots possibly implemented later on.