Thursday 13Jun19 - Research Description
I had a scare this morning. We're scheduled for another Python workshop today, and we were meant to have finished some exercises for it. I did the exercises on Tuesday night on my home computer, never thinking about where exactly that file would be saved; it was saved to my laptop, meaning that when I came in this morning, I had no access to it. I had to decide between running home and uploading it from there, and simply doing it all over again on a computer in the lab here. I decided that it seemed very manageable to do again, and it might even be a good idea to work through it once more. I was able to finish and upload it without much of a hitch, so in the end, no worries.
Yesterday I met with Rebecca Larson after lunch and we talked through the big picture aspect of the project she wrote her paper on. Showing me the poster, she explained each aspect of how the data had been collected and analyzed. She emphasized that much of the heavy lifting had already been done for her, and that her primary responsibility had been to write code to find patterns within that data once it had been documented. Hubble's GRISM (a very fine prism) was used to stare at four small areas of the Hubble field, letting in light from very specific galaxies chosen for study. The detector was rotated into five different orientations to help distinguish galaxies that might be close to each other. The GRISM spread that light into a spectrum of wavelengths, and from the visibility of the galaxy in a given spectrum, she was able to determine which galaxies were heavily redshifted. There can be some confusion about this, since a galaxy might simply look far away because it is faint. Redshifted galaxies do not appear in the ultraviolet spectrum because ultraviolet light is easily absorbed over long distances and time. The purpose of her code was also to help find spikes in several spectra at corresponding wavelengths, filtering out noise. What resulted was a fairly convincing profile of a galaxy, with information about its distance and makeup.
My project will be similar in most ways. We are looking at 12 galaxies from three view orientations for a shorter amount of time; these galaxies are less redshifted than those Rebecca studied. Rebecca's code is finished, but will need to be scaled up for the larger amount of data. I'll be selecting locations associated with those galaxies, and cross-referencing that with their spectra to determine how far, how redshifted they are, and what their makeup is. This is an exercise in cataloguing heretofore unstudied galaxies!
For now, my main goal is to keep studying Python and settle in. It's overwhelming being in a new place, working with information and techniques one is wholly unfamiliar with. Rebecca wants me to focus on learning to code as well as I can over the next few weeks, while we gently delve in to the technical aspects of my research. It's hard to avoid comparing oneself to others when research experience and skill levels are so unevenly distributed. I feel like a fish out of water. I'm grateful to my research team for giving me a long period to adapt. My primary goal here is to learn, not to unfold the secrets of the universe in two months.
The hardest bit of information to grasp yesterday was this: color, whether seen by our eyes or by a machine, is not discretely what it seems to be. Even when something appears to be blue, what our eyes are detecting is really an average of countless wavelengths around that shade of blue. To truly understand what color a thing is, and why it appears to be a certain color, the light from it must be spread out into the spectrum of wavelengths that make it up.
Yesterday I met with Rebecca Larson after lunch and we talked through the big picture aspect of the project she wrote her paper on. Showing me the poster, she explained each aspect of how the data had been collected and analyzed. She emphasized that much of the heavy lifting had already been done for her, and that her primary responsibility had been to write code to find patterns within that data once it had been documented. Hubble's GRISM (a very fine prism) was used to stare at four small areas of the Hubble field, letting in light from very specific galaxies chosen for study. The detector was rotated into five different orientations to help distinguish galaxies that might be close to each other. The GRISM spread that light into a spectrum of wavelengths, and from the visibility of the galaxy in a given spectrum, she was able to determine which galaxies were heavily redshifted. There can be some confusion about this, since a galaxy might simply look far away because it is faint. Redshifted galaxies do not appear in the ultraviolet spectrum because ultraviolet light is easily absorbed over long distances and time. The purpose of her code was also to help find spikes in several spectra at corresponding wavelengths, filtering out noise. What resulted was a fairly convincing profile of a galaxy, with information about its distance and makeup.
My project will be similar in most ways. We are looking at 12 galaxies from three view orientations for a shorter amount of time; these galaxies are less redshifted than those Rebecca studied. Rebecca's code is finished, but will need to be scaled up for the larger amount of data. I'll be selecting locations associated with those galaxies, and cross-referencing that with their spectra to determine how far, how redshifted they are, and what their makeup is. This is an exercise in cataloguing heretofore unstudied galaxies!
For now, my main goal is to keep studying Python and settle in. It's overwhelming being in a new place, working with information and techniques one is wholly unfamiliar with. Rebecca wants me to focus on learning to code as well as I can over the next few weeks, while we gently delve in to the technical aspects of my research. It's hard to avoid comparing oneself to others when research experience and skill levels are so unevenly distributed. I feel like a fish out of water. I'm grateful to my research team for giving me a long period to adapt. My primary goal here is to learn, not to unfold the secrets of the universe in two months.
The hardest bit of information to grasp yesterday was this: color, whether seen by our eyes or by a machine, is not discretely what it seems to be. Even when something appears to be blue, what our eyes are detecting is really an average of countless wavelengths around that shade of blue. To truly understand what color a thing is, and why it appears to be a certain color, the light from it must be spread out into the spectrum of wavelengths that make it up.
Comments
Post a Comment