Building Virtual Communities: What I learned from sliced
Nick
Wan is the Manager of Data Science at Kentucky Fried Chicken and
co-host of the competitive data science game show SLICED. He previously
worked at the Cincinnati Reds as a Data Scientist. He streams Twitch
where he broadcasts SLICED as well as data science projects using data
from NFL, MLB, Pro Bullriding Association, Rocket League, Overwatch
League, and NWHL. Nick Wan earned his PhD in psychology from Utah State
University.
satRday is dedicated to providing a harassment-free and inclusive conference experience for all in attendance regardless of, but not limited to, gender, sexual orientation, disabilities, physical attributes, age, ethnicity, social standing, religion or political affiliation.
We do not tolerate harassment of participants (including organisers and vendors) in any form. Sexual innuendos and imagery are not appropriate for any conference venue, including presentations.
Anyone violating these rules may be given warning or expelled from the conference (without a refund) at the discretion of the conference organisers.
Our code of conduct/anti-harassment policy can be found here.
You’re a Parselmouth, Harry! Speaking python inside of R workflows
Alec Wong
Sometimes you need to speak snake. Becoming proficient in both Python and R has become a necessary skill in today’s data analytical landscape, to maximize productivity among communities of analysts working in either or both platforms, and to take full advantage of the unique capabilities each language offers. Rstudio has written a package for the R language that unifies operation of both languages within the R session: {reticulate}. I will demonstrate an ETL procedure used in my daily work that uses python functions sourced from a package of my own design to extract data from a Microsoft OLAP cube, and embeds these functions into an R script, seamlessly integrating into a {dplyr} pipe chain. From this demonstration, R users will learn a) the basics of calling python from within R, b) how to structure the python functions to work with {dplyr} verbs, and c) common strategies for maintaining reproducibility when using both languages.
Geocoding European Soccer Stadiums
Andres Rodriguez
Use R and Google Maps Geocoding API to be able to geocode the address of European Soccer stadiums.
Scalable Shapley explanations with the fastshap package
Brandon Greenwell
I would like to present a brief overview of Shapley explanations in R, and the contributions the fastshap package brings to the table; in particular, faster and more scalable feature Shapley values for general machine learning models in R.
Using R to build models for National Football League projections
Eric Eager
At Pro Football Focus, a company based out of Cincinnati, OH, we supply data and analysis tools to all 32 National Football League teams and 100 NCAA football teams. As a part of this partnership, we provide projections for players as they transfer from NCAA to NFL. In this talk I’ll discuss how I use modeling tools in R to simulate a player’s set of future outcomes using a number of different data sources and user inputs.
R-driven Power BI
George Mount
Microsoft has developed an impressive stack of BI tools, including Power BI for dashboards and reporting. But sometimes you just need a statistical programming language to really get the job done. In this presentation, we’ll look at how to integrate these products, where R can come in handy, and demonstrate a couple of use cases along the way.
Intro to CatBoost, and its approach to categorical variables
Jay Cunningham
Tree-based models are top performers, but handling categorical variables can be a challenge. Several R packages do have support for factors, but they have different approaches and usually can’t handle an unlimited number of categories. CatBoost has a sophisticated approach based on target encoding that performs well and overcomes such limitations. I’ll briefly go over those other approaches, outline how CatBoost handles factors, and end with an example.
Survey, Linguistics, Analysis kNowledge Guide: SLANG, a tool for multi-faceted text analysis in R
Julia Haines
I have built an R Shiny App that enables the user to upload a text file of any format (CSV, PDF, TXT) and do some preliminary text analysis. The tool generates not only a table of unigrams and bigrams, indicating the phrases and words which show up with the most frequency, but also allows the user to conduct topic modeling and sentiment analysis. Sentiment Analysis is conducted based on a lexicon dictionary chosen by the user, and topic modeling is conducted based on topic definitions provided by the user. This tool can then output an HTML document that users can use without having R on their machine.
The Current State of NLP and Text-Based Machine Learning Modeling
Michael Kearney
My talk will provide a snapshot of the current tooling and package landscape for processing and analyzing natural language (text) data. Consideration will be given to open source packages in R and python, allowing for comparisons and a brief discussion of whether and how to import python modules in the R environment.
Using R To Identify SARS-CoV-2 Variants In Mixed Samples
Mike Sovic
A variety of monitoring efforts have played important roles in informing scientists and policymakers about the dynamics of the Covid-19 pandemic. As more focus has turned toward the significance that new Sars-Cov-2 strains, or variants, may have in the context of things like ease of transmission, disease severity, and vaccine efficacy, some monitoring efforts have in turn expanded to not only test for the presence of Sars-Cov-2, but also to leverage genomic sequencing technologies to document the identity of the specific variants present in a patient or location. As part of a broader monitoring program, the State of Ohio has incorporated such Sars-Cov-2 surveillance at wastewater treatment facilities, which have a unique property of being able to provide information on a large target population with a sample from a single location. However, such samples also present analytical challenges, as most software developed and used for identification of Sars-Cov-2 variants to date has focused on samples from individual patients, and is not able to effectively evaluate samples that contain mixtures of variants. In this presentation, I’ll discuss MixViR, a work-in-progress R package designed to address the challenges of detecting and identifying multiple viral variants from a single high-throughput genomic DNA sequencing sample.
“FREE!” R/Plumber APIs for all!
Pete Gordon
Have you ever wanted to push your R code out so everyone can use it over the Internet with a simple API call, and you only pay when it is used an extreme amount—which is preferably never extreme—this is for you. Google Cloud Run is an excellent tool for this. Deploy your R/Plumber API in a Docker serverless environment with no worries and make your amazing Data Science available to everyone. Let’s do it together!
How to get started with R: Tales from a former complete novice
Shane Arnett
I think there are a lot of folks out there who get involved with groups as a way to start building their expertise in R and analytics in general. I know when I started, I was completely lost and had no idea where to start; if it weren’t for the larger analytics and R community, there’s no way I would be as far along as I am now. My plan for the talk would consist of 4 main (brief) ideas: 1) a bit of my background (which had zero to do with programming) 2) How I got started and what I do currently as a senior digital analyst in insurance 3) a dump of great resources for folks that are just getting started 4) tips and tricks/things to keep in mind for the new R user or those curious about trying it out. Would likely also present a small side project I created as an inspiration for people to seek out projects that interest them in order to practice.
Building Virtual Communities: What I learned from SLICED
Nick Wan
Nick Wan is the Manager of Data Science at Kentucky Fried Chicken and co-host of the competitive data science game show SLICED. He previously worked at the Cincinnati Reds as a Data Scientist. He streams Twitch where he broadcasts SLICED as well as data science projects using data from NFL, MLB, Pro Bullriding Association, Rocket League, Overwatch League, and NWHL. Nick Wan earned his PhD in psychology from Utah State University.
