"Official" April 2019 BARUG Meetup

This is a past event

117 people went

Location image of event venue

Details

Agenda:
6:30 Pizza and Networking
7:00 Fah Sathirapongsasuti: Welcome to 23andMe
7:05 Announcements
7:10 Robert Gentleman: ALTREP: Alternate representations of basic R objects
7:25: Hui Lin: Use blogdown + Netlify to build and deploy your website
7:45: Robert Horton: Clustering of categorical features using community detection on directed co-occurrence graphs
8:15: Bri Cameron: Exploring activity patterns with HealthKit and R

************************
Hui Lin
Use blogdown + Netlify to build and deploy your website

This presentation will show you how to use the R package 'blogdown' and Netlify to build and deploy your website. We will go through the whole process from creating a simple site to having the website up and running. We will also give a brief overview of different ways to build and deploy a website using Netlify. It will be hands-on. R code examples and website templates are available on GitHub for participants to follow along on their laptops.

*************************
Robert Horton
Clustering of categorical features using community detection on directed co-occurrence graphs

It is often convenient to characterize items with sets of labels, for example, articles in a database might tagged with several keywords per article. These tags can be useful for indexing the items (as in keyword search), but they may difficult to use in machine learning, particularly if there are a large number of different tags and some of them are more or less rare. I will demonstrate several examples of constructing conditional co-occurrence graphs from tagset types of data, and how to cluster the items (or tags) using the Louvain community detection algorithm. These clusters may be useful as low-cardinality features in machine learning.

************************
Dr. Bri Cameron is a biostatistician in Data Collection at 23andME

Exploring activity patterns with HealthKit and R