Network analysis and visualization in R

This talk will discuss the use of network analysis and visualization tools in R to gain social and research insights. Network science is a rapidly growing area that allows us to understand the dynamics of interconnected systems: social, digital, physical, biological, economic, political, and others. Contemporary network research examines increasingly large and complex digital trace datasets. Seeking new and effective ways to examine and present such information has become a priority in many areas of science, journalism, and the industry. In all of those spaces, R is the platform of choice for statistical network models. Examining and visualizing networks in R provides a powerful way to deliver insights to experts, policy-makers, and lay audiences.

Katherine Ognyanova

An Assistant Professor at the School of Communication and Information at Rutgers University. Her research areas include network theory and methodology, computational social science, social use of technology, civic engagement and political behavior, social media and mass communication. She is experienced in data science and information visualization.

You can follow Katherine on Twitter as @Ognyanova.

Ten Years of Revolutions

December 7 marks 10 years of continuous publication at the Revolutions blog (blog.revolutionanalytics.com). In this talk, I’ll reflect on why the blog was first created, review some of the milestones it marked over that time, and share some statistics from the blog’s history. I’ll also review the many technological revolutions that have occurred over the past 10 years and their impact on the R community.

David Smith

David is the Cloud Developer Advocate at Microsoft. After graduating with a degree in Statistics from the University of Adelaide, South Australia, David spent four years researching statistical methodology at Lancaster University (United Kingdom), where he also developed a number of packages for the S-PLUS statistical modeling environment. David continued his association with S-PLUS at Insightful (now TIBCO Spotfire) where for more than eight years he oversaw the product management of S-PLUS and other statistical and data mining products. David is the co-author (with Bill Venables) of the tutorial manual An Introduction to R, one of the originating developers of ESS: Emacs Speaks Statistics, and a member of the board of the R Consortium.

You can follow David on Twitter as @revodavid.

Analyzing Music with spotifyr

spotifyr is an R package for pulling data from Spotify’s Web API. We’ll go over a few projects that have used it and walk through how to get started on your own music analysis.

Charlie Thompson

Charlie is a Data Scientist at Storyblocks, an online subscription service for stock media. He uses R and Shiny to analyze A/B tests and search algorithm performance. An avid music fan, he also posts analyses and Shiny apps combining data science and music on his blog at RCharlie.com.

You can follow Charlie on Twitter as @_RCharlie.

Functions: using, writing, and non-standard!!

Functions are a key component in making your code readable, modular, and sharable. Let’s write functions in R, use them, and apply them to data. The tidyverse purrr library provides a way to map functions in a pipeline. Life moves on as expected until you plan to incorporate your own functions while using tidyverse functions. All of a sudden you can no longer pass in strings from your function to the tidyverse functions in the function body because of non-standard evaluation (NSE). Let’s demystify NSE, so you can get back to coding!!

Daniel Chen

Graduate Student at Virginia Tech (VT) in Genetics, Bioinformatics, and Computational Biology (GBCB). Works at the Social and Decision Analytics Laboratory at the Biocomplexity Institute at Virginia Tech.

You can follow Daniel on Twitter as @chendaniely.

How to build a racist AI without trying

In this talk, gwynn will use sentiment analysis and build a model that exposes the ethnic biases in data sets commonly used to train algorithms. Correcting this will involve using training data from ConceptNet, an open source semantic network from MIT, as well as the reticulate package, to incorporate Python datasets into an R environment. A theme of this talk is the ethical responsibility data scientists have when developing algorithms to reduce bias.

Gwynn Sturdevant

Gwynn Sturdevant uses data to improve transportation. At Volanno, she improves estimates of flight and train arrival times. Using sentiment analysis, she analyzes twitter and facebook posts to facilitate car purchases. Prior to this, gwynn completed a post-doctoral fellow with Ken Kleinman (University of Massachusetts Amherst and Harvard University) where she developed an extensive Shiny application to assist in balancing trials. Gwynn has a PhD in Statistics from University of Auckland, the birthplace of R, where she studied under Thomas Lumley and Ross Ihaka. For her PhD, she focused on finding long-term effects, after cessation of treatments, when noisy measurements cross a threshold (blood pressure, high cholesterol, diabetes). Prior to her PhD, Gwynn taught high school math and Spanish. She has a Masters in Mathematics from Northern Arizona University, a BA from St Mary’s College of Maryland, and Diploma in Teaching (Secondary) from the University of Auckland.

You can follow Gwynn on Twitter as @nzgwynn.

How to write respectful and empathic code and documentation

You can tell a lot about a person by the code and documentation they write (or don’t write). In this talk, you’ll hear about best practices and workflows for R packages, projects, and scripts that demonstrate respect and empathy for users. Along the way, you’ll learn how your work can encourage contributors and promote the development of aspiring data scientists and budding software developers in the R community and beyond.

Martin Skarzynski

Co-Chair, Bioinformatics and Data Science Department, Foundation for the Advancement of Education in the Sciences (FAES) Martin Skarzynski, a Cancer Prevention Fellow since 2017, is passionate about Bioinformatics, Data Science, Epidemiology, and Statistical Computing. Martin uses the Python and R programming languages and command line tools to explore, analyze, visualize and present data and has a strong interest in reproducibility, scientific publishing workflows, and open data/science best practices.

Outside of the Cancer Prevention Fellowship Program, Martin is co-chair of the Bioinformatics and Data Science Department at the Foundation for the Advancement of Education in the Sciences (FAES), where he has been an instructor since 2015 and currently teaches Introduction to Python (BIOF309). Martin is also an instructor for Software and Data Carpentry, non-profit organizations that teach computational skills, and a Data Science content reviewer at Smartly Institute.

Martin holds an MPH in Epidemiologic and Biostatistical Methods for Public Health and Clinical Research and a certificate in Data Science from Johns Hopkins University, a PhD in Tumor Biology from Georgetown University, an MS in Biotechnology from Jagiellonian University in Poland, and a BA in Biology from St. Mary’s College of Maryland.

You can follow Martin on Twitter as @marskar.

Influencing Factors on High School Graduation

Using data from the US Department of Education, I examine a number of metrics that have a positive and negative correlation with graduation rates. I also develop a model to predict graduation rates and changes necessary to increase graduation rates.

Helen Levy-Myers

As a Data Scientist, Helen sees her role is to create compelling data visualizations and engaging stories. She focuses on dense visualizations with multiple avenues to enter and explore the data. Her blog has examples of her data analysis and visualizations. She works with large data sets from the EPA, CDC, US Department of Education, DC State Superintendent of Education, and the California Attorney General. She regularly posts her R code on GitHub and RPubs.

You can follow Helen on Twitter as @hlevymyers.

Spatial Data Analysis in Power BI: From Beautiful Thematic Maps to R-based Models

This presentation is aimed at people who want to create beautiful thematic maps, perform impactful spatial data analysis, and explore common R-based models with Power BI. Much of the content will make use of Power BI’s out-of-the-box spatial capabilities through Esri’s ArcGIS Maps and the MapBox Visual for Power BI.

While I use data from the United States Census Bureau and the Socrata Open Data APIs, what you’ll learn can also be reproduced using similar non-U.S. spatial data.

I’ll also highlight some commonly used R-based models you can integrate with Power BI. From clustering algorithms to hotspot analysis, R in Power BI allows us to deliver better and more rigorous insight from our spatial data.

If you’re accustomed to mapping data in R or Shiny and want to be able to do the same at scale and to a broader audience, this is the session for you.

Tony McGovern

During the day, Tony uses machine learning models and data mining techniques to solve hard, analytical problems for his clients. When it’s time to show the results of his analysis, he often turns to Power BI, Shiny, or D3.js to build expressive, interactive applications.

You can follow Tony on Twitter as @tonmcg.

How to run rstudio servers anywhere with containers - HPC, Cloud, and locally

R is the the main programming language used by the bioinformatics team at the Inova Translational Medicine Institute. Rstudio server as an IDE has helped tremendously. However, there are currently several limitations:

- the rstudio server needs to be initiated by root
- each machine can only have one server running
- the user cannot change the version of R it points to

Container solutions like docker and singularity solve all three of these issues. In this session, I will show how we run rstudio servers and ensure reproducible research using both docker and singularity on a shared server. I will also show how to make use of the free resource provided by Google Cloud Shell to have your own rstudio server for quick prototyping. Finally, I will present the issues that we are still working out for discussions.

Wendy Wong

Wendy has over 10 years of experience in international interdisciplinary research. She has in-depth knowledge in statistics, computer science and biology. She was previously a computational biologist at the Wellcome Trust Sanger Institute where she worked on genetic interaction map for Caenorhabditis elegans. As a statistician, she also worked at Winton Capital Management in London on Transaction Cost Analysis. Before joining Inova Translational Medicine Institute, she was a bioinformatics scientist at Illumina Ltd. in Cambridge, UK, working on algorithms on next-generation sequencing data. Wendy has a Ph.D. degree in Biometry (Biological Statistics) from Cornell University, where she developed statistical models to model the processes of molecular evolution using sequence data. She holds M.S. degrees in Biometry and Computer Science from Cornell University. She also has a B.S. degree in Genetics, Bacteriology, Mathematics, Computer Sciences and Statistics from University of Wisconsin - Madison.

3D mapping, plotting, and printing with rayshader

Is there ever a place for the third dimension in visualizing data? Is the 3D pie chart truly bad, or just misunderstood? In this talk, I will show you how you can create beautiful 3D maps and visualizations with the rayshader package. In addition, I will talk about the value of 3D plotting, how interactions with the R community helped drive the development of rayshader, and how writing/blogging about your projects can vastly improve your code. And, of course—lots of beautiful 3D maps and figures.

Tyler Morgan-Wall

Dr. Morgan-Wall lives in DC and is a research staff member at the Institute for Defense Analyses in Alexandria, Virginia. He has a PhD in Physics from Johns Hopkins University and is the main developer of the rayshader and skpr packages.

You can follow Tyler on Twitter as @tylermorganwall.

Animate All the Things: Methods for Animated Visualization in R

An overview and walk-through of best practice for developing data gifs to tell stories by recreating several notable visualizations and animations using the R graphics ecosystem.

Michael Lee

Michael is a former special education teacher with 5+ years of R development experience focused on the creation of products such as visualizations, animations, interactives, reports, and applications to disseminate survey research and other data knowledge.

You can follow Michael on Twitter as @mikeleeco.

Learning R by Doing

This talk is intended for beginners to R.

As a new learner of R, it can be hard to find the right resources. Not because there aren’t any, but because there are too many quality sources! There are also numerous ways to solve any given problem, making it even more difficult.

I found one of the hardest things about getting started with R was simply knowing how to find quality help and resources. Something as simple as learning the lingo to Google was infuriating. Then the overly technical pages and hundreds of resources were overwhelming and daunting.

From a pure beginners’ standpoint, I was confused by customized RStudio setups. I didn’t understand packages. Once I did, I had no idea which package was “the best.” I was utterly confused.

In this talk, I will help sort through the mess and provide resources that helped me get started. The R community is the most supportive learning community I’ve encountered. Welcome! I learned R by doing. I found an interesting dataset and dove in, teaching myself what was needed along the way. This talk will share pitfalls I encountered and resources that helped get my learning back on track.

Matthew Hendrickson

Matthew is passionate about data, analytics, and research. He has over a decade of experience constructing research projects, analyzing data, and developing policies & strategies. He has developed and deployed analytics solutions for higher education clients to increase marketing efficiency, student recruitment, and student retention. He developed and taught Data Mining to Masters students. Finally, he has provided research and analytics consulting to doctoral students and private businesses.
You can follow Matthew on Twitter as @mjhendrickson.

Time-Series Regressions - What Are They & Why Would I Use Them

It not only sounds scary, it looks scary. Whether its actually knowing if your model predicts anything, or how much hassle you’re in for converting those not-quite time-stamps to POSIXCT via lubridate, maybe, you’ll probably need to know when and how to do them. This is a introductory talk by a novice R user who had to learn the hard way but saw her project (a financial audit) come out all the better for it.

Minh Tran

Minh is originally from Los Angeles and moved to DC this past June. A Sociology major, she discovered the joy of R her junior year at UCLA through a seminar on data science by the Sociology department, and realizing how versatile it was beyond just statistical applications. Her interests pertain to issues around the environment, labor, and public policy.

You can follow Minh on Twitter as @eatdictionaries.

Understanding the International Space Station Crew Perspective Following Long Duration Missions:
Data Analytics and Visualization of Crew Feedback

This presentation will discuss how the use of an analytical framework in conjunction with the current human interface, improved the understanding of the International Space Station (ISS) crew perspective data and shortened analysis time, allowing for more informed decisions and rapid development of improvements. These data analytics and visualization methods significantly optimize valuable ISS post flight crew debrief qualitative data, yielding results that can be applied to both current and future spaceflight design and development, and other domains. This presentation will discuss a collaboration that has allowed a team of Human Factors engineers at NASA’s Johnson Space Center (JSC) to analyze and share data in a more automated and accurate fashion, thanks to the efforts of the JSC Chief Knowledge Office (CKO). Trends are no longer manually derived and are visualized effectively to assist in presenting these evolving techniques and subsequent results to an ever-growing population of human spaceflight end users.

David Meza

David Meza currently serves as the Chief Knowledge Architect at NASA Johnson Space Center (JSC). During his tenure at NASA, he has worked in all aspects of the Information Technology field developing and deploying several IT systems in use at JSC. His desire to improve IT processes and systems lead him to earn Master’s certificates in Project Management and Six Sigma in addition to becoming a NASA certified Lean Six Sigma Master Black Belt. In his current role at JSC, he established the Operational Excellence program, promoting a viewpoint of organizational leadership that stresses the application of a variety of principles, systems, and tools toward the sustainable improvement of key performance metrics by focusing on the needs of the customer, empowering employees, and optimizing existing activities in the process. Mr. Meza is conducting research on Automatic Classification algorithms, domain specific search interfaces, topic modeling, data driven visualization and a Bayesian network model for risk analysis. He holds a Master’s in Engineering Management from the University of Houston Clear Lake.

You can follow David on Twitter as @davidmeza1.

From DoD-IC Signals to Biological Signals: Cloud Scale Genomic Signal Processing for Robust Microarray Data Analysis

With the increase in the Volume, Velocity, Variety in microarray large-scale cancer genomic (LSCG) analysis there is a need to determine a “Tipping Point” when a decrease in microarray gene quality (MGQ) adversely affects accuracy & performance. This research implements a Cloud-Scale Distributed Processing (CSDP) environment for novel separable parallel one dimensional wavelet transformation in R. We also establish a microarray wavelet denoising method which encompasses:

• Providing Distribution, Parallelization, Scalability in Genomic Signal Processing
• Wavelet transformation and thresholding
• A means to gauge the robustness
• Proficient classification and parameterization

Benjamin Harvey

Dr. Harvey holds a Cryptologic Computer Science Certificate from the University of Maryland Baltimore County, a B.S. in Pre-Medicine/Computer Science from Mississippi Valley State University, a Masters in Computer Science from BSU, and a D.Sc. from BSU’s Department of Computer Science. Dr. Harvey’s accomplishments include research at Harvard-MIT Division of Health Sciences and Technology in the Bioinformatics and Integrative Genomics program in 2008. Dr. Harvey completed a Bioinformatics Post-Baccalaureate Research-Fellowship in 2009 within i2b2, National Center for Biomedical Computing, with an academic appointment at Harvard Medical School. Dr. Harvey also conducted research in 2010 at the National Institute of Health, Clinical Center within the Department of Clinical Research Informatics. Dr. Harvey currently holds a position with the Department of Defense as a data scientist where he is leading the developing large scale distributed analytics. Dr. Harvey also holds a Part-time Faculty position with the School of Engineering and Applied Science within the Department of Computer Science and the Department of Engineering Management & Systems Engineering’s joint Data Analytics Graduate Program.

First Talk

Describe 1

Second Talk

Describe 2