For lunch we met with the other interns again; while no one really talked it was something different from the quick lunches last week.
Today was a very relaxed day. Coming in to the meeting I had nothing big planned for the day. In the meeting, we discussed our progress with each project, something we do once a week. Mr. Callens informed us about a possible bike-ride on Thursday, where I could loan his bike (mine is in Cleveland, OH). Afterwards, I sat around waiting for Matt to come, so we could discuss the problems with the data set. However, he did not know either, and it wasn't until Ben e-mailed his colleague about the problem that we were able to identify it. The coordinates were mismatched with their data by 3 rows, accounting for the weird light curves that were put out. While waiting, I helped Alexa with her project by writing a simple program to subtract two values from another (she'd have to do this 2000 times by hand). I also made my own program more efficient and user-friendly, fixing any last bugs that I could spot. When I ran the new set of data through, minor improvements in the results were made, but still not what we were looking for. I'm waiting for Ben and his colleague to discuss this; until then, I do not have much to do. I believe these constant set-backs will just make the final product even more enjoyable; while it is frustrating, it makes the day more interesting whenever I hear the phrase "I think I've found the problem."
For lunch we met with the other interns again; while no one really talked it was something different from the quick lunches last week.
Today was split into two parts: a slow and boring part, and a busy part. After our meeting, which was quiet as usual, Alexa and I went up to the Astronomy Lab. Because neither Matt now Kristina were present, I didn't have anything to work on. I ended up modifying my program, making it more user-friendly. Once Matt showed up, things got more interesting. He had finally "fixed" the data set, so I was able to run the new data sets through my program, finding new bugs. Again, each run took upwards of an hour, so I only managed to get through 4 runs. The program runs at 95% now, with some minor bugs and insecurities still to be tested.
Sadly, Matt found out that the data set has even more problems, some of which cannot be fixed. Since he will need some Light Curves for his presentation on Wednesday, we can only hope that the new and corrected data shows up in time to run through my code.
For lunch, 10 of the 11 inters (Reagan is not here) met outside. I had a burger and a soda, saving the cookies for the afternoon. Overall, the push to get actual results is still there, but it gets more and more frustrating each time the data is faulty or there are seemingly invisible bugs. I guess that's how research works sometimes. However, I am looking forward to the next few days, when we will hopefully have everything working out right.
I think yesterday used up all the exciting work for today. As I have to keep my working hours below 40, I could only work for 3 and a half hours today. Those were spent on debugging the code that I added yesterday, a very frustrating but rewarding experience. As much as debugging is a pain, it gives me more ideas and makes the working program even more special. Right before lunch, the program finally worked, so I was able to test it for a little bit afterwards. For lunch we had our weekly cook-out, where I ate a hot dog and drank a Coke. Some of the interns played volleyball, so I joined. Eventually this became a soccer circle, which most likely looked embarassing. Alltogether today was relaxing way to end a very eventful week, a week I will hopefully experience again sometime in the future.
Today was a very exciting day!
The morning started off with our first staff meeting since Monday, where we were introduced to a student from Mr. Callens' school who would be shadowing us for the day. Since we had a lecture at 10:00 AM, I did not accomplish much in the Astronomy Lab in before then. The talk was held by Dr. Dube, titled "Space Weather". He focused on the importance and danger of CMU (Coronal Mass Ejections), citing the 1859 Carrington Event as an example of the possible results of another CMU. While no other CMU of similar strength has reached earth in the past 500 years, a quick online research told me there was a CMU that missed earth by nine days in 2012. That CMU could have wiped out the power grid on the entire planet, leaving humans without power for a decade. While an event like 1859 is unlikely to happen, CMU's occur randomly, and are not very predictable. Dr. Dube wants to be able to predict these eruptions in order to give the population time to prepare for a possible Apocalypse; he is aided by the Air Force, NASA, and NOAA. Space Weather can be seen as Aurora Boreali; while they are a beautiful sight, they announce that something bad is happening with earth's magnetic field. Dr. Dube also added that life on Mars would be very difficult, as Mars does not have a magnetic field to lessen the strength of CMUs (CMUs have a magnetic charge, due to their origin in sunspots).
After the talk, I worked on outreach projects (specifically the Star-Cycle powerpoint) as Matt did not have anything for me to code. For lunch, Alexa and I attended the Astronomy meeting, where we listened to NASA's presentation of the new Exoplanet, Kepler 452-b. After listening to that, Matt, Makayla, and Aviriana gave presentations on their work (REU students). Foodwise there were subs and very tasty cookies!
After the meeting, I talked with Matt about what else to code. He suggested improving the interface and making it possible to read in an array of coordinates and have the program account for possible inaccuracies of the entered coordinates. This took me 3 hours to program, adding another 160 lines of code. The program, now at 586 lines the longest I have ever written, is almost ready to run. I never assumed that Astronomy was so heavily dependent on Computer Science. In my entire time spent on the research project, 99% was coding. I did not think this'd be fun for me, but it turns out this project awoke my curiosity as well as my push to get results! Hopefully I can continue working with this program, but I am not yet sure what Matt, Ben, and Joel have planned for me.
Today was a rather short day. Due to the late trip yesterday, we were allowed to come in later than usual. I arrived at the lab at 10:30, meaning I didn't have much time before the usual Wednesday-Talk. I talked with Matt about the Light Curves, and created error-bars to make the curve more accurate. For lunch we had pizza in the CIS Reading Room, after which we listened to Dr. David Messinger's talk on remote sensing. He is the Director of the CIS, but also works on analyzing images of earth, taken by satellites such as LandSat. In the talk, he explained the process of capturing different types of images (Black-White, RGB, Multispectral, and Hyperspectral). He explained how images are stored as cubes, and we can access any wavelength in that cube (RGB images are created by placing the three wavelengths on top of each other). Then he explained how different spectra are used to analyze things in an image. For example, using 8 different colors he was able to distinguish artificial grass from healthy grass, simply by the reflectiion properties of the two materials. Finally, he focused on how the complexity of a pixel is used to analyze large images. For example, an 8 million pixel image can be analyzed by finding the "complex tiles". Complex tiles are special because they have many different colors; a city, for example, is a complex tile in an image. Using the wavelength to create a multidimensional graph for each pixel, a computer knows how complex a tiles color components are. This process makes it easy to identify the interesting regions of such large images. After the talk, I went back to the Outreach program, as Matt's and Ben's research program was at a standstill. I revised my powerpoint on Nebulas and Star Cycles, and started the write-up for that activity.
Today was the first day without a morning meeting since interns would arrive at different times throughout the day. I arrived at nine in the morning, so I could work on the research project with Matt. On Monday I had finished with displaying the offsets for each epoch's light emission, which would account for the heat interference of the telescope. However, due to no apparent reason, the offsets were larger than expected, leaving Ben, Matt, and me with a big problem. We compared out results to those recently published by one of Ben's colleagues, and we figured that our data was off. After a brief discussion before lunch, Matt and I realized it was the program's fault, which did not take the actual average offset. I fixed this in my code, which solved the issue. Using the offsets, I was able to program my first light curve in just 5 hours. Now the research project is 95% done, as the Julian dates are not yet completely finished. It is a great feeling to know your own code has analyzed thousands of previously untouched stars.
The group (interns and REU students) met at 6 to board the bus to Mees Observatory. We stopped for dinner at Amiel's Submarine Shop, were I ate a 14" sub (a very big mistake, although it tasted great). RIT paid for everything, which made the sub even better. At around 8:30 the bus made it to the observatory; however, the sky was still too bright. The observatory's director mentioned a nearby trail, so me and 6 other interns spent an hour exploring the surrounding woods. Once the sky darkened, we went on top of the observatory to look through the telescope. In addition to seeing several planets and constellations, I made friends with the rest of the interns. We talked and created a group chat, which broke down any remaining ice between us. Overall, the trip to the observatory was a great experience, something I hope I can do again very soon.
Friday again! After a short meeting, I went straight into continuing my work from the day before. I still had the issue of getting more results than actual stars that could give results, so I spent about 30 minutes going through the output, trying to find the stars that repeated answers. Once I found a couple of those stars, I checked the actual data set and realized some stars have identical coordinates and magnitudes. As my program cannot account for that, Matt told me that was okay.
I ran the program again so I could save the output as a text file, which took me to lunch. As it was a Friday, the grill was on and the volleyball net was up. During lunch, the table I was sitting at almost flipped as our weight was not balanced on both sides. After two hot dogs, we played 5v5 volleyball until 1 PM.
After volleyball, I came back to a new problem with my program. My task was to plot the standard deviations against number of occurence of each; however, my limited knowledge of Python made this task almost impossible. As I was getting close to my 40-hour work limit, I decided to put off the plot until next week, and instead work on getting the standard deviations for the 45 micron data set.
Day 2 of intensive research! During the morning staff meeting, Mr. Callens brought in donuts for everyone, which proved to be a helpful snack during the programming that took up my entire day. As I was able to get the prototype of my program to work with the Sample Data of 800 lines, I figured it was time to try the real data on the code...A rather bad idea. Since I wasn't sure how the code would handle 1000 times as much as data as the sample, it was a hit or miss situation. Sadly, the first 5 runs were all misses. As each run took upwards of 55 minutes, I was only able to test the code 5 times, making corrections after every run.
I found 7773 stars that can be accurately analyzed, which was consistent in every run. However, the program found the standard deviation for 35000+ stars, hinting at a repetition somewhere in the code. After lunch I realized there were stars with similar coordinates in the data, so I changed my program to check for more than just identical RA and Dec. This reduced the number of assigned standard deviations to 7822, still 49 too many. The next three runs did not change anything, leaving me frustrated. Only after sleeping over it did I find repetition in the actual data, something my program cannot account for.
Looking at a Hertzsprung-Russell Diagram--which displays the possible life stages of known stars--we can see that there is a path which the majority of stars follow: the Main Sequence. However, some stars, such as Red Giants, Supergiants, and Dwarfs, stray from this line. In addition to those three common branches, a fourth group called Asymptotic Giant-Branch (AGB) stars can be found on an HR diagram. Two of the nearest galaxies to earth, the Small (SMC) and Large (LMC) Magellanic Clouds, provide very valuable information on the life cycle of all types of stars.The LMC and SMC contain a number of stars that fall into the AGB category, which is why our research will focus on infrared images of these two galaxies. There are 6 sets of data available for study, 4 for the LMC and 2 for the SMC. Over the course of 4 years (2005-2007, 2012-2014), the NASA Spitzer Space Telescope has taken infrared images every two to three months. These times, called epochs, can accurately show the change in radiation emission of the hundreds of thousands of stars measured. All the numbers come together in 6 enormous data files (upwards of 100MB each), which provide a table for every epoch of every star. Using our own code and these files, we can identify those stars that fit the minimum epoch requirement of 12 epochs, and calculate the standard deviation of measured light emission for each star. Not only are those numbers important for distinguishing AGB stars from Main Sequence stars, they can also tell us how heat noise disturbed the imaging process. We can normalize the data by measuring the influence of heat noise on constant stars, which would give us very accurate magnitudes for the AGB stars. Finally, we can graph the magnitude vs. time to make a Light Curve, which is valuable in further research of these stars and AGB stars in general. Certain patterns in light emission hint at orbiting planets or the life stage of a star.