2013-11-02
2013-10-27
Office Hours with Aaron (11:30 am - 12:30 pm) Today after office hours with Aaron, I felt much more comfortable about the class and the analyzer role in the earthquake prediction collaborative assignment. First we discussed concerns over the course, around grading, around accountability, around my experience thus far. It was helpful to share with him my thoughts in person, and have a discussion around it, since a lot of the frustration in recent days had not been shared and addressed. It was also helpful to discuss the effort that we've been putting into the class, letting him know what we've been gaining from the course, and sharing our experiences thus far with regards to the assignments and working in our vertical and horizontal groups. We also touched upon expectations and what we expect out of our groupmates and how much time and effort they are putting into the assignments, and how to keep everyone accountable and contributing. We also talked about the projectory of the course, and how we would tackle future assignments in group settings. A vertical group is also a wider term, and could encompass more than one data curator, one analyzer, one visualizer, and one presenter; it doesn't need to be defined as such, and is more flexible. We discussed how to tackle working in the large horizontal groups, and Aaron said starting Tuesday in class, our class would devote more time into breaking out into horizontal groups and communicating that way, since scheduling for a large number of individuals is indeed challenging. I think this'll definitely help us communicate with each other better, be on the same page, and keep each other accountable, since the people who don't show up will be expected to catch up on their own time as well. When we dig deeper into our new assignment, we will also work more dynamically, and the groups could be broken down between smaller subsets of horizontal groups, with people from other horizontal groups as well.
Then, after a discuss for 45 minutes around the class, we spent 15 min talking about defining the analyzer role. I explained the thoughts and feelings and confusion I had around the assignment thus far, and Aaron helped us approach our horizontal group's role. We diagramed on the board and came up with a flow of stages: worry & confusion, ambiguity & questions, defined-ness, prioritize, and S.M.A.R.T goals. Breaking down the assignment together also really helped, and diagraming how our role fit in relationship to the other roles. Another large concern was around stochastic processes, and we helped ease our worries by discussing how the class was more so organized as a breadth-first than a depth-first class. Many of us likely want to dig in too deep initially, but we're really more focused on developing depth after attaining breadth. This was a really helpful perspective, and made sense since over the course of the last few weeks, we've been improving our working knowledge of visualization, GitHub, and Python.
Overall, this was an extremely helpful office hours, and I felt much less stressed around the course and the upcoming assignments.
2013-10-30
Class on Tuesday was very similar to how office hours had gone over the weekend, a bit discussing grade confusion for students who did not attend class Thursday, and the rest of the class time breaking down the complex problem. During office hours, we had diagrammed out the analyzer role, how it fit into the workflow, and we analyzed potential parts of work. In class on Tuesday, Aaron walked through all four roles, and how each group should think about breaking down the problem into manageable parts so smaller groups could work on them.
Today I've been focusing on exploring the ETAS model and magnitude-dependent automatic alarms more, as well as exploring the already available tools for seismicity analysis we could potentially use. As a first step, I've gone through Professor Stark's slides again on magnitude-dependent automatic alarms, and plan to go through sections of Luen's dissertation in the next few days. One really helpful slide was on code that we might use. The R package ETAS was the most familiar looking resource, since it fits the ETAS model to earthquake catalogs. The USGS resource seemed like it could be really helpful, but I ran into a roadblock and wasn't sure how to open the .tar file. After asking Khoa if he knew how to open .tar files, his suggestion to use tar -xzvf file.tar reminded me that running Windows probably requires a different method of downloading the files. I found a tutorial on how to unpack a tar file in windows, and also downloaded 7-zip. After successfully unpacking the tar file, the code was all in Matlab and C. I think I'll try and visit office hours to figure out the best way to tackle this new roadblock, since the other resources, including the software from Ogata require Fortran as well, which I'm not quite sure what other tools I'll need to download.
2013-11-01
Thursday's class we worked in horizontal groups, which was helpful since everyone could participate in the discussion. The analyzers worked to break down the issue, and think about how we could split up into groups and work on different tasks, but it felt largely overwhelming to tackle. We went to office hours on Friday to work on breaking down the issue, and it was helpful for about half our group to write down potential tasks to start articulating and defining the issue, and then prioritize, and eventually create a S.M.A.R.T goal for one potential task. I think our group has a better idea of how to formulate a goal for Tuesday's class and work on the assignment.