For those of you who couldn't make it Thursday, there were two great talks (as usual). Sandhya Kambhampati from Pro Publica gave a talk on "The Property Tax Divide Analysis with R" in which she explained how, after some fantastic data cleaning, Pro Publica was able to show via an analysis in R that there was something statistically unlikely in the way properties were valued in R. This talk was a great example of R in action. Then Dirk Eddebuettel gave a really nice talk on how R can be combined with other computing languages to enhance it's power. Dirk's talk started with some early history of R and ended with a discussion of the Reticulate package, announced just last week, which allows one to call Python from R. Thanks to both our speakers for two excellent talks!
In 2017, ProPublica Illinois and the Chicago Tribune published The Tax Divide, a four-part series examining deep inequities in the Cook County's property tax assessments system. The system, which touches every family in the county, was known to be highly regressive, but reporters found that the assessor’s office ignored the problem and misled the public. ProPublica Illinois data reporter Sandhya Kambhampati will talk us through the methodology they used to conduct the analysis.
Sandhya Kambhampati is a data reporter at ProPublica Illinois, where she uses statistics and databases to uncover structural issues. Previously, she worked at Correctiv and The Chronicle of Higher Education.
The story has upended the conversation surrounding the Democratic primary for Cook County assessor, which will occur on March 20th.
From the beginning, the R and S languages were designed to provide an environment for data analysis with quick and simple access to the best numerical and statistical libraries of the time. Today, that concept has blossomed and R has continued to evolve its methods for extending the language. Dirk is going to dive into these concepts with us and show us how we all can extend R.
Dirk is an active contributor to the global R community. He is an R Foundation board member and an Associate Editor for the Journal of Statistical Software. For over 20 years, he has contributed to open source software and worked in Quantitative Finance. Dirk is one of the authors of the Rcpp package, which remains the most downloaded package on the Comprehensive R Archive Network and is used to extend the performance of roughly 10% of R packages published on CRAN.