Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

[Phys-L] Re: Charts or Graphs, and other Excel stuff



BC mentioned Tufte's book.

Actually Tufte is selling _three_ big books, which he calls a "trilogy",
plus some smaller works.

Many reviewers consider Tufte a demigod and his books to be sacred scripture.

I would give a much more measured review. Tufte's heart is in the right
place: we wants people to think clearly and communicate clearly.

However, his books are more like a tour of the portait gallery, rather
than explicit lessons on how to paint a portrait.

His books are the best available, but that's not saying much, because
there isn't a lot of competition.

I'm not sure who is the intended audience for theses books:
a) Much of the advice he gives is obvious to any experienced scientist.
That's not at all surprising; to survive as a scientist you need
to be good at visualizing data.
Non-experts often stress the importance of diagrams as a way of
communicating ideas to others ... but really, the most important
diagrams you will ever make are the ones you make for yourself, to
help you think about new and challenging ideas.

b) At the other extreme, a naive student would find reading these books
to be a hard slog ... lots of disconnected examples, without as much
in the way of clear, usable, take-home messages as one might like.
(The advice to think clearly and communicate clearly is fine advice,
but you don't need three large books to say that.)

Meanwhile, some of the other advice Tufte gives falls into the category
of taste, about which reasonable persons may differ.

I'm not saying these are bad books ... just not as super-wonderful as they
are sometimes claimed to be.

I didn't learn what I know about graphs and charts and presentations by
reading Tufte; I learned it by working with other people who know about
this stuff and care about this stuff. It is always amusing to see the
look on the faces of the new grad students, or the undergrad interns,
when they see a "practice talk" for the first time. It looks to them
like I'm being savagely attacked ... and I have to explain that no,
these are my friends, and I gave the practice talk because I *want* some
vigorous criticism of my presentation. Now tonight I will rewrite my
talk, taking their suggestions into account, and tomorrow we will have
another practice talk. After a few iterations, it will turn into a pretty
darn good talk.

One good use of Tufte's books is as a source of examples that you can discuss
in class. One that is sure to get people's attention is Tufte's nice clear
plot of space-shuttle SRB o-ring failures versus temperature ... which can
be contrasted with the much-less-clear representations NASA used, with
tragic results. You can see this example at
http://www.ercb.com/feature/feature.0008.html
about halfway down the page.