Guille el Flautista

Stanley Cup Finals, Game 2

Go Ducks!!!!
What a wild experience game 2 was! Definitely the loudest, most intense hockey game I've ever been to in my life. The Ducks won 1-0 in a game they absolutely dominated physically, and now are only 2 games away from bringing the Cup to Southern California!


My dad shelled out for the tix. Lucky for me, he decided to bring me along :)


The players emerge from the locker room to a rock-concert atmosphere!

Masters!!!

DONE!!!

The semester is officially over, we all officially passed our qualifying exams, and it is now officially Party Time!!!


Party at PHI bar after quals to celebrate Elan's birthday & the end of exams!! Happy b-day E.

Gol de Maradona vs. Gol de Messi

Maradona's goal (on the left) in the World Cup '86 in México is probably the best goal ever.
On the right is Leo Messi's goal last week for Barcelona in La Copa del Rey...he's 19 years old.

CTIO skybrightness

I am officially a published scientific author:
http://arxiv.org/abs/0705.1153

This details skybrightness measurements at Cerro Tololo Inter-American Observatory for the last ~15 years. Unfortunately, it will be Hugo's last paper.
To appear in the June 2007 issue of the Publications of the Astronom. Society of the Pacific.

Of quals & spirals...

The semester is finally winding down, which is a good thing 'cause I'm pretty sure I could sleep for a week straight. It's been a very good first year of grad school, I've definitely learned a ton, determined that Stats is the right field for me, worked on a very cool project, etc., etc....but at some point there's no sense in trying to cram more stuff into your brain & your good side says 'stop working & go to bed, dumbass'.

That said, next week is finals/qualifiers week. I have to pass 2 quals and 1 final exam before I am declared a Master of Science in the field of Statistics (oooooh!). This means that although I have long-forgotten stuff I learned in September, I somehow need to remember all of the hard stuff from every class I've taken this year and spill it out onto 1 3-hour exam in theory and 1 9-hour exam in data analysis. Yes, that's right, a 9-hour exam in which I must fully analyze a dataset and write a report. Fun stuff...

Sooo... the point is that on Friday at around 1 p.m. I will be free once more to begin drowning out everything I am trying to re-learn with sweet, sweet beer. Can't wait!!

But not everything about grad school is studying & taking exams (& drinking beer). We also work on sweet-ass research projects. My sweet-ass project this semester was the study of spiral galaxies using texture analysis techniques. The project blossomed into a very nice piece of work, with good results & promises of future work and possible publication!

The basic idea is that Astronomers have only ever classified different types of galaxies by manually inspecting images. However, new sky surveys like SDSS will return images of millions of galaxies, making manual analysis of galaxies infeasible. So in my project, I investigated numerical methods that can automatically give us crucial information about a galaxy in order to classify it semi-automatically.

My particular project was to use a new technique called the curvelet transform, that can automatically extract the interesting features of a spiral galaxy. A problem with automatically studying galaxies is that there's a lot of dust and stars that reside in the line of sight between us and the galaxy. We can try to eliminate these from our view by applying the curvelet transform, which effectively ignores point-like features (such as foreground stars) & captures lines & curves (like the spiral arms of a galaxy)...see the example below.



The original image on the left is hit with the curvelet transform & our output is the image below...which has the point-like foreground stars mostly eliminated & the spiral arms intact.

So with these sweet images, we try to fit spiral arms that are defined by an 8 parameter model. Since we have already extracted the spiral arms, the fits will not be distracted by the extraneous features in the image that aren't part of the galaxy. From these fits we then get information such as a quantitative measure of how curved the galaxy is, the inclination angle of the galaxy, etc....all the stuff that is scientifically interesting. Pretty cool, eh? (Well...i think it's cool!). Now I'm gonna try to put 100s of galaxies thru this procedure to get this information for as many galaxies as possible!




Spiral model fits for 3 galaxies.