Friday, November 18, 2011

Abstraction - or making explicit connections

One of the items in the grading rubric Marianne Staempfli and I developed for concept maps is the importance of cross links in the arrangement of the concept map:
"Cross links show complex relationships between two or more distinct segments of the concept map"
It is one of the reasons why I am currently involved in a collaboration (among some others) on genome (or transposon) ecology by co-advising Brent Saylor with Ryan Gregory, and on anthropology in ecology by co-advising Ingrid Ng with Bob Jickling.

And here is another cross link I was recently struck by: "Up and down the ladder of abstraction" by Bret Victor (no screen shot, because you have to explore the site in person) -

which I think is linked to "Engineering data analysis" by Hadley Wickham.

Both advocate an explicit visual tool to explore complicated and unexpected patterns between groups of variables. The domain specific visual language advocated by Hadley Wickham, I think, implicitly provides a model to obtain the different levels of abstractions necessary for moving up and down your data iteratively in R.

Would Hadley Wickham have Bret Victor in his RSS reader? If not, I think he should, because imagine what would happen if we put these two together in the same room.


  1. This comment has been removed by the author.

  2. Wow this is cool!

    Have you seen this site before?

    Now as long as we are dreaming up people to put in a room, imagine if we added Hans Rosling in there with Victor and Wickham!

  3. yes, it is in my RSS feed. The book is on my wish list.