Home > Aunt Pythia > Aunt Pythia’s advice

Aunt Pythia’s advice

February 9, 2013

Aunt Pythia is excited by all the snow outside and has at least a couple of appointments with a sled this morning, but before she runs off she’d like to spread her words of wisdom to her good readers.

If you don’t know what you’re in for, go here for past advice columns and here for an explanation of the name Pythia, and most importantly, please submit your question at the bottom of this column!!

And… thank you for making your questions funny and/or outrageous. Extra points if your fake name is also funny before or after I shorten it into initials. For example, you could sign your letter “From A Rotten Town”. And when I say “funny” I could mean “puerile”.

From last time:


Aunt Pythia,

How do you explain your work (and its importance/relevance to the world) to laypeople? I’m interested in your answers to this question for math, for finance, and for data science.


Dear Pre-Expositor,

First, reader Mr. Exposition had this to suggest:

When non-mathematicians ask, I usually start off by describing something simple in my general area of math that has a cool real-life application. If and only if they then ask me about what I do in particular, I start breaking out the analogies and trying to give them an idea. (This gives the other person an escape valve if they wanted to be polite but don’t want to have an intense conversation.)

I’ll add a few words too. I think it helps to know a bit about the person you’re talking to. Are they wondering what math could be useful for at all? Or are they physicists? The answer is going to depend a lot on who your audience is.

Sometimes it turns out they want to be convinced that math can be interesting to someone in its own right, and why, but sometimes they might just want to knowhow the lifestyle of a mathematician is different from that of a high school teacher. I am happy to have those conversations and leave it at that. I especially love the “why is math important one” because people who ask it often answer it without my help.

If they really want to get into the details of what you think about on a daily basis, which is pretty rare, then as a data scientist I compare my approaches to something they are aware of, for example a Netflix-like recommendation system, or a Google search-like algorithm, or a finance-style trading algorithm.

If they want to talk about what I did as an academic mathematician, I talk about elementary diophantine equations and how they get increasingly difficult as you increase the degree, and if they’re still with me I talk about seeing solutions through the eyes of individual primes, and if they are still with me I talk about the local-global principle.

I don’t try to sell academic research math as important per se, just as fascinating and beautiful.

I hope that helps,

Aunt Pythia


Dear Aunt Pythia,

You seem like a very un-neurotic person. What’s your secret? Do you have personal demons? What’s your go-to strategy for when they rear their ugly heads?

Wanna-be neuroses-free

Dear Wbnf,

I do of course have personal demons, as everybody does. I often find myself waking up at 3am thinking about things I’m behind on or things I wish had gone better. I have two pieces of advice for this kind of thing.

First, use suppression. I think suppression has a bad name. People think of it as a bad thing. They say stuff like, “oh you’re just suppressed” like that’s a crime.

But I say, use suppression to your advantage! If you can’t fix something that’s bothering you, agree to ignore it (which is an agreement you make with yourself, so nobody can even complain about it). And I don’t mean ignore it forever, either. Just make a plan to start thinking about it if and when you might have control over it. OWN your suppression and it will give back to you.

So for example, if you are stressing about your kid getting into a good kindergarten in New York City, then do what you can in terms of looking up schools and applying to them, and then after that, start up the suppression motors til you hear back. There’s absolutely nothing you can do in the meantime except fret, and you have better things to do with your time. Suppression is your friend!

Second, be pro-active. I know that’s a trite, overused phrase, but there may not be another word that means what I want to say – namely, do your best, to the best of your knowledge, on whatever it is, and forgive yourself in advance if that wasn’t enough. Of couse sometimes it wasn’t, and you have to live with the consequences, and sometimes you take notes on what would have been better. That’s ok, because the third thing is you gotta forgive yourself. It’s so obvious I won’t even make it a separate thing.

In my experience, being pro-active about something in advance, followed by 100% suppression mode, works a lot better than constantly putting something off and feeling guilty about it.

By the way, one more thing. I also let things slide. If I can’t get myself into enough of a froth to be pro-active about something, then I just let it go and I don’t look back (I do this via suppression, see above). It’s important to know when to do that too.

I hope that helps!



Dear Aunt Pythia,

I’ve decided to leave academia and become a research scientist in the tech world. In addition to my area of math, I know a bit of programming and machine learning. What else can I learn in the next few months to better prepare myself?

Rambling On

Dear Rambling,

Great question, but I’m not sure how many “research scientist” positions there are in the tech world. Most of them don’t want you to research, they want you to model! So I’m going to assume you meant something like “data scientist” if that’s ok.

First, learn python, for reals. Next, learn statistics, enough so you can explain to anyone what statistical significance is and mean it. Then, read the book I’m writing with Rachel Schutt, Doing Data Science. Oh wait, it’s not out yet. So for now, read the notes I took on Rachel’s Columbia Data Science class last semester.

And to test your new knowledge, implement the recommendation system using python. And send me the code! We’d love to have it for the book, thanks.

Good luck,

Auntie P


Now it’s time for you guys to help me answer a question. I’ve got a juicy one for you:

Dear Aunt Pythia,

As a graduate student, I enjoy attending departmental teas, if only because it’s an excuse to get away from the books for a few minutes. However, my department recently started having some of our teas sponsored by a trading firm. As somebody who has concerns about the finance industry, I am bothered by this. I thought about dumping all the tea in one of the fountains on campus, but I’d like to find a more constructive approach. Any suggestions?

Tea Party Patriot


Please please please submit questions, thanks!

Categories: Aunt Pythia
  1. c.gutierrez
    February 9, 2013 at 10:08 am

    Dear Aunt Pythia — There is some Python starter code for recommendation engines in Toby Segaran’s Programming Collective Intelligence, if that helps. signed — Cousin Hypatia


  2. February 9, 2013 at 12:04 pm

    Out the corner of my brain, I assumed Aunt Pythia was actual identity of Pythagorus of the many theories.


  3. February 10, 2013 at 9:34 am

    I hate to contradict Aunt Pythia, but I would suggest learning R instead of or in addition to Python. One of the places to start with R is http://www.burns-stat.com/documents/tutorials/impatient-r/


  1. No trackbacks yet.
Comments are closed.
%d bloggers like this: