Fleeting/Random Thoughts

[Dason]

Calculus through multiple integration is pretty much required. Working with joint distributions requires multiple integration and any program that touches any theory will require that. Linear algebra is needed when you get into the theory for linear models so it's pretty important as well. Technically if you hadn't taken linear algebra before the course I took you theoretically could get through it since it started from the basics and worked up (but we went through the basics very fast so if you hadn't dealt with them before it would be a shock). Real analysis is useful for mathematical maturity for a masters but probably isn't *required*. If you're going into a program that touches on measure theory, though, you will definitely want to have seen real analysis before that course (although I knew a few people that hadn't taken it before grad school but took it during the summer before they took measure theory).

 

[Buckeye]

My program has a theory course where we studied probability and statistical inference over two semesters. I have a strong math background, but we never really focused on by-hand integration/calculus beyond a bivariate distribution. As far as courses that I have benefited from, linear algebra is probably my number one. Matrices are all over the place. I was not much of a computer programmer before last year...I code a lot. Especially when cleaning the data set. Which is like 90% of the problem.

 

[hlsmith]

If your talking MS level, i believe up to 2 years calculus and linear algebra base at upper tier programs. But many times things can be a little negotiable. For my stats light degrees, they wanted a minimum 1 year calculus, but nobody must have looked too hard for them in my transcripts, I had taken a couple of biostats programs with good marks. Phd programs usually require a chunk of the MS work to be done with high marks before exceptance. If a person has high marks in an undergrad program and took some upper level courses they may probationary enter a PhD program i would imagine. Not, my field but I would image the venerated program comes into play as well.
Most phd programs may have a qualifying exam, but it is not require until after core course.

good buckeye and reason replied - they definitely know more than I do!!

 

[ondansetron]

I asked because it's interesting to me to see some programs say they want XYZ and I feel as thought that the bare minimum to pass through the program (maybe with a bit of self study for those students) but to really be successful and get the most from it, it would imagine almost a full bs or MS in mathematics would allow you to really kick butt. Many of the statisticians I know have at least 1 mathematics degree so it made me wonder about the disconnect.

@hlsmith, I think you're right that it depends on the school and past coursework. It probably helped that you previously handled biostats courses to back up your application.

 

[Dason]

TIL: Windows 10 - do a a three finger swipe up on the touchpad and you're brought to a window selector (which doesn't sound much better than just hitting alt-tab) but!!!!! there is an option to switch to a new desktop. How did I not know that Windows 10 offered virtual desktops? This changes everything.

So that's how I discovered that Windows 10 supports virtual desktops (which is something I've always loved in ... every window manager in Linux that I've every used) but there are apparently a lot nicer ways to interact with the desktops that I didn't realize: https://www.addictivetips.com/windows-tips/windows-10-virtual-desktop-shortcuts/

 

[Dason]

I always forget about the HAVING clause in t-sql. I've been taking to writing two step CTE where I calculate what I want to filter on in the first part and then just apply the filter using WHERE on the previous result. This can be useful sometimes when I want to pull in other info or do other related stuff - but a lot of times a simple HAVING clause would get me what I'm looking for.

 

[ondansetron]

Windows defender is now a strong, competitor in the free antivirus market. Only problem is it is incredibly protective and often blocks things it shouldn’t. I’ve had it 2 weeks now and am constantly adding exceptions to the program list.