This search:shape coding - ResearchIndex document query at CiteSeer brings up a healthy page of citations of a paper I wrote back in 1990. It was published in the Proceedings of the first Conference on Scientific Visualization sponsored by the IEEE, NSF, and NASA.
Shape coding of multidimensional data on a microcomputer displayis the link to the actual article on the ACM site. It declares that there are 16 other papers that cite it in their primary bibliography or text.
The abstract from the publication:
"The visual representation of data from complex systems, whether databases, measured scientific data, or simulation output, holds the promise of discovering patterns in the data that will increase its management efficiency while revealing relationships invisible to numeric methods. In this paper we present a simple and flexible method of shape coding for higher dimensional data sets that allows the database operator or discipline scientist quick access to promising patterns within and among records or samples. The example used is a thirteen parameter set of solar wind, magnetosphere, and ground observation data collected hourly for twenty-one days in 1976. "
The story in a nutshell:
In the late 1970's I was doing design studies based on quilt patterns as a graphical design exercise, and it jogged a memory. A few years earlier I had worked on a statistical research project tracking population movement in Rhode Island during the 1890s, and the quantity of data was overwhelming. The primary investigator on the project commented that I had an unusual knack for looking at pages of data and "seeing" patterns in the numbers without actually going through them item by item. At the time it occurred to me that a computer program that translated the numbers into visual patterns as I used them would be a good invention. The quilt patterns seemed a good vehicle for organizing this complex data.
I started thinking about this and doing some research. It seemed like a good enough idea that I actually went to night school for 13 months to get certified as a computer programmer by 1982. I immediately started writing software to develop graphic patterns from data. This had nothing to do with my actual graphic design job, as I was still working with press-on type and sable lettering brushes at the county.
A local engineer was interested in my work and offered me access to one of the first microcomputers with decent graphic display...an 8 color, 300x600 Texas Instruments pc. I spent evenings and weekends building a program for design and pattern management. I even spun it off as a graphics program and sold it through TI's third party support system as Micropix. But my real interest wasn't pretty designs, it was data-driven displays of complex, multivariate and multidimensional data.
While I was doing this, I was thinking of the use for it within organizations like county governments. In a democracy, governance was decisions made by citizens, who were rarely informed of the organizations they were making decisions about. At the time it seemed as though governments took a protective stance of hiding data in order to deflect citizen inquiry, by and large. An automated reporting system that depicted the status of all government functions according to commonly agreed upon criteria seemed like a good idea, and a good fit with the ideal of democracy.
As I made progress in feeding data into the display, I had several interesting experiences. At one point I arranged a demo to the head of the county IT department. He took a quick look and turned away, saying no manager in the county could understand what they were looking at in a pure visual display.
On the other hand, the McGraw Hill corporation flew me out to NY shortly after the Challenger shuttle disaster, and I was able to explain, with the aid of my software, how the actual votes of all 5000 of the NASA engineers could have been taken into account on the day of the launch, instead of the disasterous summary report which resulted in the deaths of American astronauts. This resulted in an invitation to work with California Testing Bureau, which was responsible for standardized student testing at over half the school districts in the nation. I worked to their specifications for over a year, as usual in my spare time on evenings and weekends, and delivered several example reports using my technology on their data. The outcome of that experience was experience, not cash, as they continued to express interest in my ideas but didn't feel the market was ready for a product based on them yet.
By this time, around 1988, I had contacted NASA, specifically the Godard Space Science Data Center. Dr. Lloyd Treinish, the head of the Data Center, invited me to attend the first Scientific Visualization conference at JPL in 1988. I scraped together some cash and flew out there. The encounter resulted in me standing up in a room full of scientists asserting that there were ways to visualize multidimensional data effectively that had not surfaced in the literature for various reasons. While many of the younger scientists argued that it would be in the literature if it could be done, a senior researcher from NASA Ames who was tapped to chair the next conference invited me to organize and chair a workshop at his conference. The topic was to be the visualization of multidimensional data.
Val Watson, the senior Ames researcher, endured my countless questions and enthusiasms with great forebearance. I called most of the major corporate research centers, universities, and supercomputing centers around the country tracking down leads to people who might be doing research in this area. I ended up inviting about a dozen presenters, all of whom accepted the invitation. The Conference was held at Stanford University in February of 1990. One of the presenters I had invited, Larry xxxxx, was head of the Naval Resarch Labs computer facilities, and took me somewhat under his wing. He grilled me on my approach, and talked me up to his peers. In the meantime I chaired the two day workshop, and we had presentations from Bell Labs, the Santa Fe Institute, IBM, Godard, and numerous other institutions.
This resulted in an invitation to put my work into a paper and submit it to the follow on Visualization conference to be held in San Francisco that June. The paper was accepted, and I was invited to create a workshop on Multi Dimensional visualization for that conference, too. Eventually I chaired workshops at the 90, 91, and 92 conferences, delivered a tutorial at the 92 conference, and was on the steering committee for the series of conferences.
In 91 the conference was in San Diego, and I was approached by the head of the Lawrence Livermore Labs computing facility and asked if I wanted to participate in a research project. The project was the International Thermonuclear Energy Reactor project, ITER, which was the first international research project involving Russia, the US and Japan after Glasnost. My role was to devise visualizations of the engineering design points for the proposed reactor under the direction of the head of Information Sciences for the project, a Dr. John Perkins, from England.
Following the wearying pattern of doing this all in evenings and weekends, I worked part time for Livermore for about 8 months, producing a series of visualizations and co-authoring two papers for the project staff meetings.
ITER was not a priority for the US after the first Iraq war, and my sponsor on the project pulled back from outside contracts. Eventually the US involvement in the project ground to a halt. I have heard recently that it might be starting to gain some momentum again.
By late '92 it was clear that my marriage was headed for dissolution, I had run out of energy and support for my work.
While maintaining interest, and co-chairing the panels committee for the 93 conference with Dr. Treinish, I opted out of attending the Conference, and dropped my involvement in the research. Technology had outpaced me, my limited macintosh and pc access at home was complicated by advances in programming and operating system technology that rendered my skills somewhat obsolete.
Things sat dormant for a long time, with occasional contact with the people I had met through my Conference activities.
By 2002 I had upgraded my home computer equipment and invested in a development system. I wanted to program a simple graphics utility for home use that I could sell and help meet my oldest son's college expenses. As I acquired proficiency in the new environment, I began to see ways that I could recode the old visualization routines into an even more powerful tool.
In spring and summer of 2003 I spent an intensive few months bringing the old technology to life in a new platform. I got it to the point where I could depict all 3000+ counties in the US in one display, with point detail for 40 census parameters for each county depicted simultaneously: some 120000 data points color and shape coded for visual inspection. A subordinate display on the same screen depicted the individual county under the cursor on the main display, with detailed info on that county, and a browsable surface that showed the distribution of values for the entire data set for any given parameter.
At that point I decided I needed a support system for this kind of work. My efforts to bring this to the county's attention again met with no interest. Recently this might have changed.
Posts
