LA R Users: May Meeting

LA R Users Group will have our May meeting online at https://usc.zoom.us/j/92828963940

Talk1 (20min talk, 10min Q&A)
Speaker: Rich Iannone, Software Engineer at RStudio PBC
Title: Making Tables with the gt Package
Abstract: The gt package is great for making beautiful tables using tibble or data frame data. Its interface is designed to be both straightforward yet powerful. There are simple functions for those everyday display table needs but also plenty of options for customizability should you need it.

There is a defined table model that subdivides a table into its component parts. These include the table header, the stub, the column labels and spanner column labels, the table body, and the table footer. All of these parts (and their component cells) can be targeted for the application of styling or the addition of footnotes. With all of the available options, there are many ways to customize the look of your table.

The best way to get a feel for what gt can do is to take it for a spin, and so we’ll do that through some live demos. I’ll walk you though some code that progressively adds components to a gt table. Table cells will be formatted and, through styling, their appearance will be radically changed. The result in each case will be an HTML table that can be used in a variety of data products.

Talk2 (25min talk, 5min Q&A)
Speaker: Avik Das, Software Engineer at LinkedIn
Title: Dynamic programming for machine learning: Hidden Markov Models
Abstract: A Hidden Markov Model deals with inferring the state of a system given some unreliable or ambiguous observations from that system. One important characteristic of this system is the state of the system evolves over time, producing a sequence of observations along the way. By incorporating some domain-specific knowledge, it’s possible to take the observations and work backwards to a maximally plausible ground truth.

This talk explores Hidden Markov Models in three steps:

First, I define Hidden Markov Models and how they apply to machine learning problems.

Next, I build up an understanding of the Viterbi algorithm, used to infer the state of the system given a sequence of observations. This involves some basic math, but the goal is to form an intuition for the algorithm. Some sample Python code is presented to demonstrate how simple the algorithm is.

Finally, I introduce several real-world applications of Hidden Markov Models in machine learning. In this section, real-world considerations like feature extraction and training are discussed.

Basic math knowledge is expected, just the ability to express concepts as equations and an understanding of Big-O notation. Basic Python knowledge is helpful, but not required. The goal is build up intuition.

Code of Conduct: https://github.com/laRusers/codeofconduct

Schedule
6:15 open zoom
6:30-7:00 First talk
7:00-7:30 Second talk
~8:00 Virtual Social

LA R Users Group
Invite yourself to our Slack group: https://socalrug.herokuapp.com/
Ask us any questions by email: larusers@gmail.com
Find our previous talks on GitHub: https://github.com/laRusers/presentations
Follow us on Twitter: @la_Rusers
Check out more events: https://laocr.org/

***** DON'T BE SHY *******
Reach out if you want to be our speaker. First-time speakers are welcome