Week+of+10-1+to+10-7

October 1
Today Grace and I arrived at 3:45. We couldn't do much until John came. When he did, we had a little meeting with Grace, John, Dr. L and I about where we should be headed. It turns out we have a decent amount more of exploring to do in ManyEyes, and then we will look at histograms in the e-lab to find various pieces of evidence as to where the cosmic rays are in the data. Dr. L showed Grace and I the use of highlighting points in ManyEyes and showed us the functionality of highlighting many points and changing the axes and seeing where the points move. We also learned that if you roll the mouse over a point, the coordinates of that point (as well as its event number) are displayed. I started into my ManyEyes fiddling by highlighting a group of points that I concluded last week were probably cosmic rays. I first highlighted all the points of the highest mass and changed both axes around to various combinations and concluded that this group of points did not make much of a pattern in any of the graphs. I decided to try to hone in a little more on the characteristics of what I thought were the cosmic rays. I plotted eta 2 vs. mass and highlighted the points with a mass above 60 GeV (60 is about where the trend around 0 begins to become clear) that had an eta 2 within 0.2 of 0. I decided to see how many of these points also had an eta 1 within 0.2 of 0. I was hoping that I could eventually isolate the particles that were near vertical and of high mass, two key components of cosmic rays as detected by the CMS detector. The graph on the left below is the graph of eta 2 vs mass with my choice of points near eta 2 of 0 highlighted, and the graph on the right is eta 1 vs mass with the same points highlighted. Most of the points with an eta 2 within 0.2 of 0 also had an eta 1 within 0.2 of 0. I eliminated the points from my set that had an eta 1 outside of 0.2 from 0. It turns out that while I was doing this, Grace took a similar approach to looking at the graphs and made an interesting discovery. She highlighted a series of points across all masses that surrounded eta 2 of 0 and then switched to eta 1. The graphs are below with eta 2 vs mass on the left and eta 1 vs mass on the right.

This method showed an interesting trend. The particles with eta 2 around 0 //and// mass above 40 GeV are the ones that stayed grouped around eta 1 of 0. So, there is a certain mass cutoff point that keeps the trend intact. I had just estimated it by guessing it was around 60 GeV, but Grace actually found the trend. This makes sense that there would be a mass cutoff between the massive group of small particles scattered around all different etas and those of high mass which were concentrated around 0. The reason this is not terribly surprising is that we already established that in the CMS detector mass essentially equals energy, and the cosmic rays are of a certain high energy and thus high mass. It appears that 40 GeV is about the energy where particles intentionally generated by the experiment become sparse and cosmic rays are the majority of what is left. I then adjusted my data set accordingly, and honed it down to the particles of mass above 40 GeV which have both eta 1 and eta 2 within 0.2 of 0. The graph below is eta 1 vs eta 2 with my group of highlighted points surrounding the origin. Now that I knew I had a set of points that were clearly of mass above 40 GeV and of eta 1 and 2 within 0.2 of 0, I was confident I had a group of cosmic rays, or at least a group containing cosmic rays. With this solid data set, I decided to explore into the phi graphs and see if cosmic rays showed a pattern on the phi axis. Before doing so, I called John over and had him clarify exactly what eta and phi are, because I was feeling a little fuzzy and didn't want to botch my analysis. Once he re-explained, I realized I had been missing something. I realized that values around 0 for eta 1 and eta 2 did not ensure that the particles were traveling vertically, only that they had passed through the full diameter of the cylinder. They could have passed through it on any phi angle on the plane of eta 0. To illustrate this, I took a couple pictures on the 3D view of the CMS detector. This first one shows the perspective from which eta 1 and 2 are measured. They would be measured away from the vertical line in the center of this snapshot. Thus, any eta measurement actually contains an entire plane that is a slice across the diameter of the detector that looks something like this following picture. This entire circular plane is contained in an eta value, so a data point with eta 0 could actually be a particle that went laterally across the detector, for example in a horizontal line across the center of this circle. This would not be consistent with a cosmic ray, so we need phi in order to further narrow down our search. One note to remember when looking at the perspective from which phi is measured (the circular slice above) is that phi does not start from vertical, it starts from horizontal. So, the vertical angle which we expect cosmic rays to enter on would be phi 1 of -pi/2 (about -1.57) or phi 2 of pi/2 (about 1.57). Given this newly clarified information, I graphed phi 1 vs mass and phi 2 vs mass and, with the exception of a few outliers, found exactly what I was looking for. Pictures of the graphs are shown below (phi 1 on the top, phi 2 on the bottom). After discovering these patterns, I am extremely confident that my group of yellow dots consists mostly of cosmic rays. After I made my confident conclusion, it was time to leave at 5:45.

October 3
Today I arrived at 4:00 and John had us start in on the CMS e-lab graphing tool. We looked at the dimuon data set from 2011. We each had it display the graphs of eta, phi, and mass. We started learning how to use the tool with John's help, and toyed around with some different settings. The e-lab graphs are histograms, so they resemble bar graphs rather than scatter plots like ManyEyes. The y-axis is always the number of events and the x-axis is one of the various measured values, so a histogram is specifically designed for the analysis of larger scale patterns by showing how many events in the set returned a specific value of a certain quantity. John taught us about the bin width, and how it represents the width of the bars on the histogram. For example, in the graph below, the x-axis goes from -0.2 to 0.2, and the bin width is set to 0.1, so there are 4 bars of width 0.1 to span the graph. Two other functions of the e-lab tool shown by the above graph are zoom and cut. Each of these functions can only be used after you click and drag to highlight a certain area of the graph. The zoom button is purely visual, and just lets you zoom closer on a certain part of the graph. The cut button, in addition to zooming, also adjusts the data set for all the graphs to only include the points in the highlighted area. So, below I have shown the phi graph before and after the cut I made to the eta graph above. This last phi graph is actually very significant in that it supports the conclusion I made when using ManyEyes. I concluded after using ManyEyes that cosmic rays were particles that had eta near 0, phi's near -pi/2 and pi/2, and had a mass of 40 GeV or greater. To test this, I cut the section of the mass graph from 40 to the far right edge, and the picture below is the resulting phi graph. This graph made the trend even clearer. Isolating the points that were near an eta of 0 and above a mass of 40 GeV created a clear trend in the phi graph around -pi/2 and pi/2, supporting my theory on where the cosmic rays appeared in the data. After the success in the e-lab tool, Dr. L came in and spent a while discussing various aspects of our plan moving forward in the class. After this I began to plan my part in the group project about cosmic rays, and I left QuarkNet at 6:00.