After attending the useR! 2016 R Conference, I felt supercharged and armed with new insights and ideas about how to further contribute to the R community. I met wonderful people in real life (IRL) from twitter and heard interesting case studies about using R for large data analysis. I already have two R packages on Github but wanted to expereince the detailed (yet rewarding) process of submitting a package to CRAN (Comprehensive R Archive Network). For my first package, I had looked to what my interests were and what was needed in the community, which turns out to be more analysis-ready datasets about beer.

Yes, BEER!

Beer analytics is the use and analysis of data to gain insight about breweries, production, operations and its impact. The U.S. Department of Treasury has a voluminous collection of historical data and reports about beer statisitcs as reported by breweries at the National level. In my previous analysis project, about web scraping craft brewery data in Connecticiut, I noticed the minimal amount of data that was available for beer-centric analyses and more interestingly the data that was publicly available on the data.gov was locked in PDF’s or static HTML tables. This problem presented the perfect opportunity for me to contribute a dataset package and learn more about package development.

The process for creating the R package was fairly well documented in the R Packages book by Hadley Wickham. I also looked for guidance by exaiming existing R dataset packages, like janeaustenr by Julia Silge. The process to obtain the first dataset on beer materials was a horrible tale of copy and pasting 10 years worth of data from PDF format reports, but once that was was complete I was able to web scrape 8 additional datasets on the historical tax rates of beer, wine, and tobbaco products. (look for those new datasets on the development version 1.1.1, until I submit to CRAN in early August)

My first submission for review, I made a silly mistake and did not upload my package to win-builder for checking the source code on a Windows machine - I will never make that mistake again. My reviewer, Kurt, was prompt and courteous in informing me that an error arose from a faulty site certificate in a supplied url in my documentation from https://www.ttb.gov/beer/beer-stats.shtml. After I fixed this error, I re-submitted and was waiting by my email with schoolchild-like wonderment until my acceptance into the nearly 8000 R package community was confirmed. I’m really honored to have taken this important step from just being a consumer of R packages, to a R package developer. The R community has been really great in their feedback via Twitter:

Here is my first package available on CRAN: ttbbeer, which I will continue to update with new “liberated” datasets and use this as an opportunity to increase education and advocate for analysis-ready datasets on open government data portals. Here is the wiki to follow along on the journey to more Open Beer Data. Cheers :beers:

To install the package type the following:

install.packages("ttbbeer")
library("ttbbeer")

Or you can install the development version from Github:

library(devtools)
install_github("jasdumas/ttbbeer")
library(ttbbeer)