Another week done at Galvanize! It’s been a full week and a busy weekend, so I’m posting a bit later than I’d like. Also my jokes are getting nerdier (see post’s title) though I’m OK with that. This week was all about…

Photo credit: LendingMemo

…Statistics! I was excited about this since I’ve had some stats classes in the past. My classes had generally focused on frequentist statistics, which is just one approach to statistical thinking. I’d previously learned basic inferential statistics and hypothesis testing, so I was (nerdily) most excited to learn more about Bayesian statistics. I’ve mostly approached statistics within the context of public health research or drug development, so thinking about the parallels to web traffic, for example was new to me.

Some of the highlights for me this week were learning about bootstrapping, getting some practice with Bayesian methods and calculating Bayesian posteriors, and learning about different approaches to A/B testing. The final assignment of the week was to work on the “multi-armed bandit” problem, where we explored different ways that one might test versions of a webpage to determine which generates the most clicks. It’s pretty cool to see how you can set up an experiment that will automatically converge on the version with the best click through rate… assuming you’ve implemented your algorithms correctly.

I’ll also point out that each week, we work on 5 different individual assignments and 5 pair assignments, where we practice pair programming with another person in our cohort. This has been a great way to 1.) meet everyone in our cohort! and 2.) learn from each other. Each day, you might end up with someone who knows more or less about the particular topic than you, or sometimes the pair is pretty equally matched. Either way, you definitely learn a lot from each other! I’ve enjoyed this aspect of the program, especially knowing that pair programming is a popular approach in industry.