A Summer Full of d(mv)/dt *d
Return to the Summer Journals forum
Login to post to this thread
Summer Retrospective - Aug 25, 2008 at 12:58AM | |
---|---|
Paul Abbazia 12 Posts |
This summer was certainly an adventure. I made new friends and saw new places. I even got to work on a pretty cool project. I lived smack dab in the middle of ever-so-busy Washington, DC. I worked at NASA with a group of very intelligent people. Perhaps the most exciting part was to be completely on my own entering a new life experience. I go to a local college with about a dozen or so of my high school class mates, so I've always had a cushion of familiarity to fall back on. This turned out not to be a problem; I made some great friends and had a blast during my SPS summer. |
Week of July 28, 2008 - Aug 5, 2008 at 5:42PM | |
---|---|
Paul Abbazia 12 Posts |
What was a restful weekend led into a very busy week. |
Week of July 28, 2008 - Aug 1, 2008 at 4:24PM | |
---|---|
Paul Abbazia 12 Posts |
The internship is rapidly approaching its end, along with a wonderful summer in DC. This week, we all gave our presentations at the American Center for Physics and, for Daniel and I, we haven't had much chance to work on our research this week. |
Week of July 21, 2008 - Jul 25, 2008 at 6:04PM | |
---|---|
Paul Abbazia 12 Posts |
Work has been somewhat frustrating this past week. I've been trying a variety of things to try to find a relationship between the neutron defined cluster maps I've made and the 6 previously existing element concentration maps of Mars. I've tried finding mathematical correlations between the data sets, as well as just visually comparing the data. Additionally, I've done some statistical analysis on the clusters per element map compared to the rest of the data set. One such analysis is the T test, which is similar to computing how many standard deviations two data sets are separated by. It tests if the means of two data sets are far enough apart to be considered statistically different. Due to the possibility of error within the measurements, two data sets could have different computed average values, yet have a strong possibility of having identical averages in actuality. The other test is the F test, which tests is the variances of the two data sets are far apart or not. In the F test, it is preferable to have similar variances between the data sets, so that you know that there isn't a particularly high or low value in one data set throwing off the results. |
Week of July 14, 2008 - Jul 24, 2008 at 1:16PM | |
---|---|
Paul Abbazia 12 Posts |
I've been trying to find correlations between neutron and gamma ray data. Gamma rays have already been used to map out geochemical compositions on Mars, however neutron use has been limited to identifying hydrogen/water. This has involved a lot of image comparisons, along with some statistics. One such statistic is the Pearson correlation coefficient, which measures the degree that two variables correlate, ranging from -1 to 1. 1 is complete correlation, -1 is complete inverse correlation and 0 is no correlation. Unfortunately, this hasn't revealed any strong correlations. |
Week of July 7th - Jul 11, 2008 at 5:06PM | |
---|---|
Paul Abbazia 12 Posts |
This past week has been busy. I've put together cluster maps of the neutron data, along with the combined GRS + neutron data. I've started analyzing the neutron data for relationships with gamma ray data and we are going to experiment with other clustering software soon. |
Week of June 30th - Jul 7, 2008 at 10:11AM | |
---|---|
Paul Abbazia 12 Posts |
Currently, I'm in the process of performing clustering analysis on the Martian neutron data as well. We're hoping to see correlations between the neutron and gamma ray data so that we have the possibility of using neutron measurements to do geochemical analysis in the future. |
Week of June 23, 2008 - Jul 1, 2008 at 4:45PM | |
---|---|
Paul Abbazia 12 Posts |
I spent this week trying to replicate previous work done on Mars Odyssey gamma ray spectrometer data. This work involved k-means cluster analysis, a similar type of analysis to PCA. K-means cluster analysis attempts to find groupings in the data. I was able to create a color overlay map for the surface of Mars, indicating which sections of Mars lay within which groups of the (normalized) data. |
Week of June 16, 2008 - Jul 1, 2008 at 4:44PM | |
---|---|
Paul Abbazia 12 Posts |
I spent this week reviewing the differences between multiple PCA programs, as well as attempting to analyze the Mars Odyssey elemental concentration data using PCA. The data has many invalid entries in it, and I wrote a program to remove the invalid entries. Additionally, I did a few sample plots using Goddard's PCA program. |
Week of June 9, 2008 - Jul 1, 2008 at 4:43PM | |
---|---|
Paul Abbazia 12 Posts |
This week at NASA: I'm now up to two textbooks from work, gamma-ray astrophysics and radiation detection and measurement. This week has had the start of us getting used to some of the work we're going to be doing. I did some atmospheric gamma ray absorption modeling. Additionally, a professor from Rowan (my school) arrived this week. Professor Klassen is getting acquainted with the project, and the majority of his research has focused on Principal Component Analysis (PCA). PCA just happens to be what Daniel and I will be working on next week. PCA takes a coordinate system and attempts to find a relation between the base vectors that define the system. It then creates a new coordinate systems defined by vectors among the lines of greatest variance. For example, suppose you have 4 lists of data, each representing a different element over a series of locations. Your initial coordinate system is defined by say the Fe vector, Si vector, H2O vector, and B vector. PCA may output four vectors in decreasing variance, with say Fe-Si as the vector of greatest variance, then Fe-H2O, then Si-H2O, then Fe-B. It may be found that the contributions of Si-H2O and Fe-B have a much smaller relevance to the data, so the system reduces to two base vectors, Fe-SI and Fe-H2O. This simplifies analysis and allows classification by the most statistically significant relations. We're to implement a PCA algorithm in Interactive Data Language (IDL) and, as a learning exercise, replicate the results of PCA analysis from Mars Odyssey. Non-work related activities: More traffic hijinx. Firetrucks have pinwheels on the front. Some stop lights in DC have "Do not run this red" signs. Attempting to circumscribe the city to avoid traffic only results in running into more traffic. And this morning several blocks of stoplights had their power cut. Surprisingly, traffic seemed to flow better, as fear of crashing seems to be effective at curbing aggressive driving. The groups of cars would approach and stop at each intersection. Whichever side had a car move first would then stampede across the intersection, until a driver from another direction was brave enough to creep out into traffic. Dan and I went to a ballroom dancing class with some of our coworkers at NASA. We felt out of place, but it was fun. Jack Trombka (one of our bosses) retired this week, only to take up a new position at NASA. We did the proper intern thing and took home the left over party trays. GW's gym is tiny and cramped. Kunal didn't care for it, but I'm getting decent use out of it. We both run just about every day (usually separately), and it's a good way to see some of the sites in town, as well as to get a feel for the layout of the city. Post edited July 1, 2008 at 4:44 PM EST. |
Week of June 2, 2008 - Jul 1, 2008 at 4:38PM | |
---|---|
Paul Abbazia 12 Posts |
As the first week of this internship comes to a close, I'm having a great time. I'm really enjoying the internship so far. All the information I've had to absorb about LEND this week makes me glad I've finished most of my physics courses, quite a bit of topics are touched upon. Even still, I'd like to have a stronger background in chemistry, astrophysics, and imaging for this project, but I'm very glad that I at least have the background that I do. I've met a lot of other interns here in DC. Most of them can be described as follows:
Everyone is DC is really passionate about something. Many about politics, Goddard employees about space, and my fellow interns about Physics. It's kind of strange to see so many motivated people. All the other interns are really nice. I'm having a good time, and we have a lot of fun. |
Introduction - Jul 1, 2008 at 4:02PM | |
---|---|
Paul Abbazia 12 Posts |
Hey there! |