R Lighting Talks for March 2020 Meetings

Cleveland R User Group

1. Pick a data set from below
2. Do something interesting with it:

• create some visualizations that tell a story about the data
• do an exploratory analysis of the data
• build a predictive model (machine learning)

Data Source 1

RStudio Learning Survey Data

github.com/rstudio/learning-r-survey

RStudio has conducted informal surveys in 2018 and 2019 on how R users learn and use the language. Carl Howe presented the 2018 results in his rstudio::conf 2019 talk The next million R users. The 2019 results have not been officially analyzed and released.

All the results are available in the repository rstudio/learning-r-survey. The 2018 results are the safest option because they have already been analyzed. I (John B) was able to import the 2019 results with readr::read_tsv(), but there were some parsing errors, so this could be more complicated.

url2018 <- "https://raw.githubusercontent.com/rstudio/learning-r-survey/master/2018/data/survey_English.tsv"
survey2018 <- read.delim(url2018, stringsAsFactors = FALSE)

url2019 <- "https://raw.githubusercontent.com/rstudio/learning-r-survey/master/2019/data/2019%20English%20R%20Community%20Survey%20Responses.tsv"
survey2019 <- readr::read_tsv(url2019)

The full text for the 2018 survey questions is in Learning R Internet Survey - Question Names.tsv. The full text for the 2019 survey questions is in survey-questions-2019-en.csv.

Data Source 2

Kaggle Housing Prices practice competition

kaggle.com/c/house-prices-advanced-regression-techniques

Contribution Guide

The lighting talks will all be presented from one computer to reduce the transition time between talks. Crucially, no code will be executed for the presentation. In other words, in addition to your source code (e.g. R or Rmd file(s)), you will need to submit the finished product to display (recommendations below).

If you are comfortable with Git and GitHub, please submit a Pull Request to this repository with your contribution. Commit all your contributions in the submissions directory in a subdirectory titled with your first and last name, e.g. firstname_lastname. Please do not let the complexity of Git/GitHub discourage you from contributing a lightning talk. Feel free to email Tim with your contribution, and he will add it to the repository.

Below are recommendations for what to submit based on the output of your analysis:

1. Plot(s) - Submit the plot in a web-friendly format such as PNG or JPEG. From the RStudio plots pane, you can click Export->Save as image... to export a PNG file. Alternatively you can use png(), jpeg(), or ggsave() directly in R. Also include the R script(s) you used to generate the plots.

2. Reproducible Report - If you used knitr/rmarkdown to generate a reproducible report of your analysis, submit the R Markdown source file and also a Markdown version of the report. Markdown is preferred because GitHub will automatically display the Markdown; whereas, it doesn't do this for HTML and other formats. For best results, use the output format github_document().

3. Shiny App - If you develop a Shiny app, you will need to deploy it yourself, e.g. at shinyapps.io. Submit the R files you used to create the app as well as a README file with the URL to your deployed app.

Questions?

Contact Tim Hoolihan