The author has a long-standing interest in data visualization for engineering design stemming from their work transforming large data sets into effective design decisions. Over ten years ago, they began creating their own visualization tools and methods out of necessity. More recently, they found insights from reading Edward Tufte on why intuitive methods for visualization work well and the importance of design. The author seeks to exchange ideas and find existing visualization methods from other domains that could be useful for their work analyzing gas turbine engine data.
1. My interest in data visualization for robust engineering design:
For some time now I have been interested in how information evolves from a
data collection phase (or learning phase) into the final design (decision phase).
Transforming data taken into effective and timely decision made, especially for
complex systems, requires good data visualization methods and tools.
I have been working with large data sets since I left the University. More than ten
years ago now (out of necessity), I started creating software tools and methods
for data visualization in my field. Two years ago, hoping for new inspiration and
to maybe learn some new tricks, I read Edward Tufte for the first time. Rather
than leaning much new I was taken by how many of “his” ideas I was already
using as they were seemingly intuitive to me. I had already been using such
methods for many years however the rational from Tufte for why these methods
work was quite insightful as I had never thought deeply about why I did what I did
and why it worked, I just knew it was the correct way to do things. Finding a
science behind it was exciting. Noteworthy as well was the reaffirmation that
design was as important as the data itself. As Steve Jobs said, “Design matters.”
Of course any good idea you have has probably been thought of by somebody
else. When reading Daniel Dennett last year, I was humbly reminded that this is
a natural phenomenon and can be understood as a convergence in design space
towards the good ideas or necessary tricks.
I have not found nor will I find, on my own, all the good ideas for data
visualization in the area I am working (gas turbine engines). I fully expect to find
good tricks outside my area. That’s why I am interested in exchanging ideas,
methods, book recommendations etc... with anyone on the same journey. For
example, last year I “stumbled” on a useful alternative way of displaying contour
plots for axisymmetric geometries, which increases the information to Ink ratio [E.
Tufte] substantially. The coming years will surely bring some exciting changes in
data utilization and visualization (look at what Hans Rosling is doing). William
Gibson stated it best when he said “The future is already here — it's just not very
evenly distributed”. I believe the majority of the work to be done in data
visualization entails finding the already existing ideas, packaging/combining them
in new ways and distributing them.