Getting Out our 4-D Glasses

This sixth week of research has challenged us in new and exciting ways, the most important of which is the challenge of visualizing data that has more than three dimensions.

The ATom data, and atmospheric chemistry data in general, is most informative when several dimensions can be related at once. The data calls for us to visualize each point in 4-dimensional or even 5-dimensional space.

To help us visualize such an abstract concept, we began by plotting our data in 3-dimensional space. By plotting latitude, longitude, and altitude, we obtained a 3D representation of the ATom flight path. We then colored the points by some other variable, such as dew point. This technique helped us greatly in visualizing our data in four dimensions.

Having produced a variety of these plots, we moved on to making density cluster trees from data with more than two dimensions. The coding in R for these trees is currently nonfunctional, however, and we are still working hard on analyzing higher-dimensional data with density cluster trees.

In the meantime, Cade was able to create a variety of useful time series. Incorporating the time of day into our analysis of ATom is essential to correctly interpreting the other variables; ozone concentrations, troposphere and stratosphere fluctuations, and many other important variables can change depending on the time of day.

Overall, this week marked the project's transition from two dimensions to higher, more information-rich dimensions. We conclude this week with optimism and excitement to discover what higher-dimension analysis will tell us!