What a week!

The phrase "Just in Time Learning" is oft repeated at the IAA. When I was teaching, we called it "generating need". The most effective learning happens when a student is presented with a problem, and recognizes that they can't solve the problem unless they learn a specific new skill. When students know what they need to know, education isn't a one-way, teacher-to-student endeavor. The concept sticks because it is so wanted. It's like finding the long-lost, missing piece of a giant puzzle, and finally being able to complete the picture.

This past week, the ultimate example of "just in time" learning occurred for me. Throughout the Fall, Dr Healey has been broadening our Python horizons, little-by-little. And just this Friday afternoon, we took a look at some Git materials curated by John Jernigan. The two converged Friday night after a package I had installed was simply NOT WORKING. After 30 minutes of googling, I found ONE page where the author of the package described what was causing the error, noted that version delivered from pip install would throw the error, and directed readers to get the updated, yet-unreleased version directly from his GitHub. Were it not for the incremental increases in Python knowledge, and Jernigan's "need to know" git knowledge-base, I would have had NO IDEA how to move past the error.

This "just in time" method plays out so well at the IAA that I have a hard time believing it's intentional. (Side poll -- what % of his actions would Healey say are intentional?) But then again, our faculty are pros at what they do. Somehow it all comes together when it needs to come together. Which is, I suppose, why IAA alumni tell us, #TrustTheProcess.

Incidentally, we are exposed to so much material, it is easy to think we are making no progress. Indeed, it is a year of FFTs, and there is always something new. Looking into even the recent past is a great way to see how much change has really taken place. My experience this week makes me think that the "% of stuff I understand on Stack Overflow" metric is the most reliable indicator of how much we are actually learning at the IAA. Classmates, try it out! Pick a post and marvel at how much of it would be incomprehensible to your May 2020 self.

Let's quickly take a look back at this week!

• Dr John Elder of Elder Research presented an incredible Analytics In Practice lecture on the top three things he's learned about doing data science. Most fascinating to me is the concept of target shuffling. Anyone intrigued by the problem of research reproducibility and publication bias should definitely look into it!
• Healey continues to dazzle us as we start working with Tableau and SAS Visual Analytics. (But, yes, that is an MS Paint graphic above. Worthless confidence interval added for good measure.)
• Lead Ethan-Cole Evans and Scrum-Master Elizabeth Toller are bringing their Agile planning experience to bear on our practicum team activities. Shout-out to Yeimi Chavez-Gonzalez for checking delivered data for PII and Victor Cardeno for crushing it in the tech-lead role.
• Gus, Shalina, Andrew, Luna and I finally settled on a model after going several rounds with atmospheric data and SAS's proc arima. Dr Susan Simmons was NOT kidding about the complexities of fitting ARIMA models!

Next week promises to keep our hair on fire. The need for a bird's eye view eventually became too large to ignore, and I tried my hand at making a pseudo-Gantt chart. The overview shows at least six team assignments or projects in the works over the next three weeks, as well as a host of individual day-to-day obligations. #TrustTheProcess!