DataPhilly August 2013 - Data Science with R

Hey Everyone,

Our August meetup will be focusing on R (http://www.r-project.org/) and related technologies.

Agenda:

6:30 PM - 7:00 PM - Food, Networking, and a word from our sponsor

7:00 PM - 7:30 PM - RStudio (http://www.rstudio.com/) - A Tool for Reproducible Research, by Oliver Will

7:30 PM - 8:00 PM - Visualizing Census Tables, by Richard M Heiberger

8:00 PM - 8:30 PM - Lightning Talks

8:30 PM - Leave for Nodding Head (https://plus.google.com/117164681594927807176/)

More Details:

RStudio as a Tool for Reproducible Research, by Oliver Will

Abstract

I often use R at my job. Now I always use it through the RStudio IDE (Integrated Development Environment). I will introduce RStudio and its features in this talk. RStudio was designed to integrate with tools that enable reproducible research, unlike generic IDEs such as Eclipse that were designed to manage large and complex program builds. I'll cover managing your R Scripts with Git for version control and making html reports with Markdown and knitr through RStudio. As time permits, I might show how to connect to an R instance on Amazon Web Services. Finally, one major feature RStudio lacks is a debugger, and I might address that as well.

Bio

Oliver is a Senior Manager at IMS Health in Plymouth Meeting. He has a PhD in Applied Mathematics from the University of Southern California. (Go Matt Barkley!) He has taught statistics at colleges and has worked as a statistician at consulting and market research firms for the past 10 years. Once, he contributed a package to Bioconductor, which no one downloaded and is no longer there.

Visualizing Census Tables, by Richard M. Heiberger

Abstract

The U.S. Census Bureau traditionally summarizes data by cross-tabulated tables. Tables contain a lot of useful information. Tabular structure hides complex relationships. I have been working with a group to design graphics that communicate information to the user better than tables. We illustrate with several sets of Census tables along with our recommended graphs.

One of the graphical forms we use is the diverging stacked bar chart. This graphical technique is appropriate for rating scales, such as Likert scales, which are very common in marketing research, customer satisfaction studies and other fields. Likert scales consist of choices from the sequence: "Strongly Disagree" < "Disagree" < "No Opinion" < "Agree" < "Strongly Agree". I will discuss implementation of an R function for diverging stacked bar charts. I will show a wide set of applications for this technique, including several applications with Census data.

Bio

I have been an R and S-Plus user since 1988. I am Professor of Statistics Emeritus at Temple University and teach the Graduate Statistical Computing course. My current consulting practice focuses on designing easy-to-read graphics for complex data situations.

Meetup Location Details

We'll be meeting on the 16th floor. You'll need to sign in at the front desk. See you there!

Directions to Nodding Head

Head South down 15th St (past City Hall)

Walk 4 blocks, turn right onto Sansom St

Nodding Head is on your left (230 ft)