Last week, I attended a Tokyo-based Statistical Software R meetup named “TokyoR #45” held on Jan 17 at VOYAGE GROUP office in Shibuya, Tokyo. Almost all presentations were given in Japanese, but in this post I’ll share brief a summary of those presentations in English.
The meetup consists of 3 sections. Beginner sessions, Advanced sessions and Lightning Talks (LT).
Nobuaki Oshiro (@doradora09) gave a three minutes version of his “Learning R in 10 minutes” presentation which includes what is R, who should use R, how to install R, how to code R and where to acquire information about R on the web.
Tetsuro Ito (@tetsuroito) talked about “Hot topics of R in 2015”. He introduced some packages, such as “anomalydetection” by Twitter, “ver 1.0 of ggplot2” by Hadley Wickham. Also he stated nowadays most of hot packages are stored in github, not CRAN. Additionally, he introduced the newly released book “The Lean Analytics“, a methodology for data scientists.
Shinya Uryu (@u_ribo)’s presentation was “Data pretreatment for Data pretreatment”. According to his presentation, to reduce data pretreatment time enables expanding time for data analysis. That makes us get high-quality output. Also, he recommended to use R project file (.RProj) and R markdown file (.Rmd) on the Rstudio that integrate team members and their deliverables.
Yatsuta Toshihisa (@tyatsuta) talked about “Typed Function”. As you may know, to process huge size of dataset on R requires too long CPU time, however the “Typed Function” enables fast processing with efficient memory allocation. The “Typed Function” is 250 times faster than normal loop.
Yohei Sato (@yokkuns) talked about Kernel-Multivariate analysis. As you can see the background of his slides are a Colonel Sanders. Actually, the character of the KFC “Colonel Sanders” is known as “Kernel Ojisan” in Japanese.
Lightning Talks (Short presentations)
Yoshio Tokorosawa (@dichika); also known as Serial Package Creator (Seripac) developed an project schedule management system on R. His system utilize some R packages sinchokuR, AnomalyDetection and twitteR. SinchokuR retrieves the schedule data from github and AnomalyDetection checks is schedule behind or not. When the system caught behind schedule, then notify it to developer via twitter. Also, he announced he is translating the book “Advanced R” with some folks and it will be released in Japanese.
K Mori (@wonder_zone) developed a favorite anime character recommendation system with SVM (Support Vector Machine) using dplyr and e1071 packages. Source code of his system are available at his github repository.
I, Takekatsu Hiramura (@hiratake55) talked about the newly released book “R and cloud computing” written by Ajay Ohri and myself. I introduced some cloud service providers which are compatible with R such as Amazon Web Service, Google Prediction API, BigML, Microsoft Azure ML, plot.ly and Yhat.
Tatsuya Tojima (@salinger001101) shared an idea of using the continuous integration tool Jenkins as an analytical reporting tool. His project makes daily reports using R and Jenkins on batch processing automatically.
There were more than 80 R users from software, consulting, banking, social media and other industries. They shared idea, talked and drank a lot.
Next meetup TokyoR #46 was scheduled on Feb 21. When you have any opportunities to visit Tokyo, please join our meetup. If you have any questions or requests please feel free to contact me.