Form, Fabrication, and Data Environments
with Brian Brush and Yong Ju Lee of SoftRigid (2009)
featured in Kottas, Dmitris. Contemporary Digital Architecture: Design and Techniques Barcelona: Links, 2010.
In the Spring of 2009, Brian Brush, Yong Ju Lee, and I embarked on an independent study investigating the extent to which a spatial dataset may inform the programming, design, and fabrication of an object. The scale represented by the dataset happened to be much larger than that of the desired object. The project takes raw data, analyzes this information through GIS mapping, and uses the analysis as the basis for a formal design. At the end of the semester, a collection of dynamic and interactive maps was presented along with a full-scale prototype of the designed object. For me, this was an exercise to learn Flash, explore a few representational opportunities for animated maps, play with automating procedures in ArcGIS, and find analysis methods that may contribute to SoftRigid's design process.
THE DATASET (POINT)
The dataset contained the number of connections made to 270 wireless routers distributed around Columbia's Morningside Campus, per hour for the week of final exams for the Fall 2008 semester (7-13 December 2008). In total, the database contained 168 (7 days x 24 hours) values per router.
Limitations include the fact that the campus, in fact, contains more than 270 routers. Givin the parameters of the research, only those routers with immediately identifiable locations (based on the name of the router) were used in the sample. Thus some buildings are not included in the research. Due to regular failures and testing, not every router recorded connection data each hour.
Ultimately, this study is research concerning a methodology, experimenting with ways of working with GIS and design and fabrication, and not rigorous research into wireless usage on campus. As a result, the dataset was deemed more than sufficient for our purposes.
USERS PER BUILDING (POLYGON)
Due diligence dictates that we start with the basics, aggregating the router data per building to get the lay of the data landscape.
The representative images included here depict the data at 14:00 on 7 December 2008, 5:00 on 8 December 2008, and 22:00 December 2008, respectively.
DENSITY MAPPING (SURFACE)
Density analyses were performed to reflect the usage of the wireless networks. The pertinent search radii based on estimates of the routers' effective reach areas.
These analyses were 2D. They were later represented in both two and three dimensions.
CONTOUR MAPPING (POLYLINE)
The density studies were converted to contours, at varying intervals. Because they were based on the density analysis, the contours were first determined in two dimensions and later represented in three dimensions.
A QUICK VIDEO
The maps generated in this analysis are summarized in the animation below.
Several thousand points, polylines, and surfaces were handed off the Brian and Yong Ju for further analysis and eventually to be used as the interpretative basis for the design of an intervening object. Together, we looked at what the data implies about wireless usage on campus in an effort to program the object, which ultimately became a seat, a place to enable mobile computing in the open areas of campus. These areas are reached by the routers but see little use. (Admittedly, our dataset reflects the winter. Again, I invoke the "testing-a-methodology" defense.)
Additional photographs of the finished prototype can be seen here.
The laptop exhibited with the bench/seat presented an interactive Flash animation resulting from the GIS analysis described above and allowing a user to dive into the data with more specificity. (Buildings could be clicked, revealing graphs of wireless connections within the user by hour for the week.) Additionally, the application allowed users presented an investigation of Butler Library similar to that done campus wide. In the case of Butler, users could see the changes in wireless connection, floor by floor and room by room. Still images of the animation are featured on the right and in the Some Maps gallery. The (very heavy and, therefore, slow-loading) Flash is here.
This project would not have been possible without the help, support, guidance, and teaching of Sarah Williams and Phil Anzalone at Columbia's Graduate School of Architecture, Planning, and Preservation. I would also like to think Brian Brush and Yong Ju Lee for their willingness to enter into this strange collaboration between fabrication techniques and the methods of GIS.