by Joseph Rickert
Last Friday and Saturday the NY R Conference briefly lit up Manhattan's Union Square neighborhood as the center of the R world. You may have caught some of the glow on twitter. Jared Lander, volunteers from the New York Open Statistical Programming Meetup along with the staff at Workbench (the conference venue) set the bar pretty darn high for a first time conference.
The list of speakers was impressive (a couple of the presentations approached the sublime), the venue was bright and upscale, the food was good, and some of the best talks ran way over the time limit but somehow the clock slowed down to sync to the schedule.
But the best part of the conference was the vibe! It was a sweet brew of competency, cooperation and fun. The crowd, clearly out to enjoy themselves, provided whatever lift the speakers needed to be at the top of their game. For example, when near the very end of the second day Stefan Karpinsky's PC just "up and died" as he was about to start his Julia to R demo the crowd hung in there with him and Stefan managed an engaging, ad lib, no visuals 20 minute talk. It was also uncanny how the talks seemed to be arranged in just the right order. Mike Dewar, a data scientist with the New York Times, gave the opening presentation which featured some really imaginative and impressive data visualizations that wowed the audience. But Bryan Lewis stole back the thunder, and the applause, later in the morning when as part of his presentation on htmlwidgets he reproduced Dewar's finale viz with mushroom data.
The slides from all of the presentations have yet to be posted on the NY R Conference website. So, all I can do here today is to provide an opportunity sample drawn from postings I have managed to find scattered about the web. Here are Winston Chang's talk on Dashboarding with Shiny, Jared Lander's talk on Making R Go Faster and Bigger, Wes McKinney's talk on Data Frames, my talk on Reproducibility with the checkpoint package, and Max Richman's talk on R for Survey Analysis.
For the rest of the presentations, we will have to wait for the slides to become available on the conference site. There is a lot to look forward to: Vivian Peng's presentation on Storytelling and Data Visualization will be worth multiple viewings and you will not want to miss Hilary Parker's hilarious "grand slam" talk on Reproducible Analysis in Production featuring explainR and complainR. But for sure, look for Andrew Gelman's talk: But When You Call Me A Bayesian I Know I'm Not the Only One. Gelman delivered what was possibly the best technical talk ever, but we will have to wait for the conference video to reassess that.
Was Gelman's talk really the best ever, or was it just the magic of his delivery and the mood of the audience that made it seem so? Either way, I'm glad I was there.