Now let's look at some actual applications
of fractal dimension in the real world.
I'm sure you've always wanted to know
the dimension of cauliflower.
Fortunately, it's been calculated
at approximately 2.8.
That is, if you look at a cross-section
of cauliflower,
this flat thing,
you find that it's actually a little more than 2 dimensional.
It's between 2 and 3 dimensional,
due to the dense fractal self-similarity
of the cauliflower branches.
You may have noticed that
the logistic map bifurcation diagram
has a tree like structure
which indeed is fractal.
So for instance, if we blow up a
little part of it here.
And some parts of it are quite
self similar.
We blow up this part here, we see that
this looks very much like the whole thing.
And people have calculated the
fractal dimension
to be about .5.
It has so many holes in it
that it's not even quite 1 dimensional.
The fractal dimension of coastlines
has been calculated.
Notice that the west coast of Great Britain
has a higher dimension than
the smoother coast of Australia.
or the even smoother coast of South Africa.
And all of these are a little bit
more than 1 dimensional.
If you look at them as going along a curve,
sort of like the Koch curve.
People have looked at fractal dimension
of more abstract kinds of phenomena,
like stock prices.
This is from a paper that was published in 2000
looking at the Oslo Stock Exchange
and looking at an index which listed
a hundred day daily price records,
a hundred week weekly price records,
and a hundred month monthly price records.
And you can see that all of them
have very similar kinds of ups and downs,
even though they're at very different
time scales.
The person who wrote this paper asked the question,
"Are stock prices following a random walk?"
And the project here was to compare
the fractal dimension of these curves
with the fractal dimension of a random walk
to see if the dimensions were the same.
If you look at random walks,
they also have a lot of detailed ups and downs
and self similar structure.
But after some complicated mathematics
this paper was able to answer "No."
The fractal dimensions are not the same.
Therefore, it's not likely the stock prices
are following a random walk.
Of course, in these real world examples there's no exact
proportion of size reduction or clear cut number of copies.
It's very different from the exact mathematical fractals
like the Koch curve and the Sierpinski triangle.
So there's a lot of caveats in actually applying
fractal analysis to time series.
A group of scientists looked at fractal dimension
as applied to Jackson Pollock's drip paintings.
They looked at the fractal dimension in these paintings
and found that if you plot the year of the work
versus their measurement of its fractal dimension
you get this kind of increase,
and this is claimed to be the evolution
of complexity
as measured by fractal dimension
of Pollock's paintings over time.
Using measures like fractal dimension
for quantifying aspects of art has
been a controversial subject for a long time.
Later on we'll talk to John Rundell,
a geophysicist who's been interested in using ideas
from fractal geometry for a long time
in the natural sciences.