David MIller


Four Dimensional, real-time data exploration tool


Introduction to Computational Media

This Processing program creates a 3-D surface plot from an external data feed (CSV spreadsheet) which can be then be manipulated to observe different aspects of the data. The ability to view the plot from different points makes it possible to draw multiple conclusions--you see different

Looking at the NYISO grid data, I saw that it is best visualized as a surface, with the ability to look at the data in both two and three dimensional views, each of which presents a different facet. Each day is broken down into 24 hour blocks, and a year is therefore composed of 365 stripes. Considering the data in one way, you see how the grid load increases and decreases across a day. Looked at another way, you see how the load increases during the summer, and at its lowest in the late fall and early spring.

You can create a surface plot in Excel or another spreadsheet, but not in real time. I'm interested in getting a datastream in real time, for example from my Tweet-A-Watt, and having the tool plot the data in real time. For a show project, that's not quite as useful as a premade dataset.

The target audience is professional data analysts and the general public--the tool is simple enough to use for anyone with a rudimentary understanding of what the visualization represents, but powerful enough to see rich meaning.

User Scenario
A user can view the data in an isometric view to get an overview (see the surface) or can use a predefined on-axis view which reduces the three dimensional surface to a two dimensional (X-Y,X-Z,Y-Z) plot-- or can navigate around the surface plot to get a better understanding of the data under study.

Using the large touchscreen at the ITP show will allow for a more immersive and potentially intuitive navigation experience, based on the degree of control afforded.

The sketch parses a spreadsheet (CSV) and separates it into individual blocks representing an hour of grid load. Each block is placed into a one-day stripe (24 blocks), and stripes are plotted sequentially. The view has an overlay providing a view of quantitative information about the data.

Programming gets much easier the more you do it. Things which seem incredibly difficult early on become easier with practice and greater knowledge. Processing libraries contributed by developers (Peasycam, ControP5) make programming the interface much easier and faster.

Using a touchscreen is a bit of an adventure but