PyData Atlanta Recap: July 2016

Previously I posted a little bit about the PyData Atlanta meetup in June, where I presented a lightning talk on my project: git-pandas. The July meetup was this past Wednesday, and it was as always very interesting so I thought I would do a quick recap of what all was presented.


The main talk this time was by Mark Guzdial, a Professor from Georgia Tech's school of interactive computing.  The talk, and his research, was focused on the problems and opportunities present in computer science education.   To start with, he made the case that not only is computer science useful, and a part of a modern liberal education, but in fact we have some societal obligation to ensure that all people are computing-literate, lest they be left ignorant to the concepts that are increasingly influencing every aspect daily life.

If this is accepted, then a bunch of interesting research questions arise from that obligation to computing education:

  • How do we correct the massive imbalance in access to early-stage CS education?
  • If we make a basic CS requirement for highschools, what happens to the bottom 10% of students? Is it worth some individuals with cognitive disabilities not getting diplomas?
  • How can CS teachers grow as teachers when there are so few of them to form a meaningful community? (~2500 across the nation in high schools)
  • If CS teachers learn enough about CS, how do you keep them from going into industry?
  • Can a teacher that isn't a programmer teach CS classes well?
  • How much knowledge of CS concepts is 'enough' for a non-CS professional?
  • Can you learn CS concepts in a meaningful way without learning a programming language?

He introduced a lot of really interesting research questions, told some stories of studies trying to find answers to them, and proposed some ideas that could really help out.

My first takeaway was that requiring a CS class for all college students is a great starting point, because then at the very least all teachers will have one class.  My second takeaway was really just thankfulness that I was fortunate enough to have good CS coursework available to me in middle and highschool.  Outside of AP Computer Science we had classes in java, visual basic, independent studies, and more, and it really did shape my college and grad school experiences later on.

Lightning Talks

David Nicholson: How I spent my summer scipy vacation

In his lightning talk, David gave a recap of his favorite talks from the scipy conference in Austin.


Ben Leathers: Predicting the Distribution of Undies: An Undie-Rated Problem

Ben talked about his experience predicting how many pairs of underwear to send to different cities for a huge fundraising bar crawl run.


Will McGinnis: Encoding categorical variables with categorical encoders

I talked a little bit about categorical encoding, which I've posted about here a few times.


Peter Molnar: First steps using Spark (and PySpark) on your desktop

Peter gave a lightning talk on PySpark, and why it's sometimes worth it to just use pyspark locally from the get-go, even though you don't need the scale yet, because it's pretty easy to and it gives you some room to grow.

Link to github page


Will has a background in Mechanical Engineering from Auburn, but mostly just writes software now. He was the first employee at Predikto, and is currently building out the premiere platform for predictive maintenance in heavy industry there as Chief Scientist. When not working on that, he is generally working on something related to python, data science or cycling.

One Comment

Leave a Reply