by this November and I was fortunate enough to have time to attend. During the conference I took notes on the presentations, which I’ll pass along to you.
Thoughts on the future of R in industry
The EARL conference started with a panel discussion on R. Moderated by Doug Ashton from Mango Solutions, the panel included Julia Silge with Stack Overflow, David Smith with Microsoft, and Joe Cheng with Rstudio.
Topics ranged from the tidyverse, R certification, R as a general purpose language, Shiny, and the future of R. I captured some interesting quotes from the panel members:
Regarding R certification, David Smith points out that certifications are important for big companies trying to filter employment applications. He mentions that certification is a minimum bar for some HR departments. Julia Silge mentions that Stack Overflow has found certifications to be less influential in the hiring process.
R as a general purpose language: Joe Cheng feels that R is useful for more than just statistics, but that Rstudio isn’t interested in developing general purpose tools. There was discussion around Python as the “second best” language for a lot of applications, and an agreement that R should remain focused on data science.
Most interesting was the discussion regarding the future of R. Julia Silge points out that Stack Overflow data shows R growing fast year over year — at about the same rate as Python. There are a lot of new users and packages need to take that into account.
I learned more about Natural Language Processing
Tim Oldfield introduces this conference as NOT a code-based conference. However, . I will mention she does a great job of summarizing how and why to perform text mining. Like all good tech, you can easily scratch the surface of text mining in fifteen minutes. A thorough understanding requires years of hard research. If you'd like an introduction to her work, take a look at her paper
I gained an understanding of machine-learning
)
found herself forced to incorporate Excel with her data workflow. Now — we all have opinions about Excel in data science — but Eina points out that for multidisciplinary data science teams, it’s good for data storage, modeling, and reports. There are issues about reproducibility and source control and for that, R is a good solution. But Eina summarizes that excel is still useful. Not all projects can move away from it.
Data science teams without structured, intentional collaboration leak knowledge and waste resources
with Starbucks revealed some of the tools used by the coffee giant, specifically (an R package for risk evaluation). Dave points out that R is an elegant language for data tasks. It has an ecosystem of tools for simulations and statistics, making risk evaluation a plug-and-play process.
Personally, I don’t have call for risk evaluation. But it’s interesting to get a quick peek into the language and concerns of this specialty.
I was reminded of the Science in Data Science
.
SOCIAL SHARE CARD GENERATOR