Thesis Meeting

I met with Kim and Chris today for the first time since my proposal defense. Since then, I’ve completed a good, working version of my data pipeline. I’ve been quite pleased with my pipeline, but I really wanted to discuss my method for determining how to extract spectra with Kim and Chris. I had just kind of made it up over a few days, and wasn’t sure if it was a good method.

I illustrated how the pipeline choses the break points for spectra by defining count-rate bins based on 25% increments around the average overall XTE count rate. Both Kim and Chris thought that it looked like a good method! (Chris’ words, I believe, were “Actually, that’s quite clever.”) One of the good things about this method is that while it does smooth out some small variations to get a bigger signal, it retains large variations. It has the potential to probe time-scales of a few days. If we wanted to probe smaller time scales, I’d need to use a different method, but this is sufficient for the goals of my research.

Things I need to do:

• figure out how to make smart_fit do 1-sigma errors instead of 3-sigma (actually, I need both)
• choose a few representative sources, and play around with the minimum time for the spectra
• plot figures using 1-sigma errors, and do some fits to constants to see what we come up with
• work on plotting a “stack” plot automatically with ppgplot