Wellington R-Users Group August Meeting
Wednesday 16 August 6:00pm – 7.00pm
Ministry of Business, Innovation and Employment (MBIE)
15 Stout Street
Ground Floor, Room G.03
Attention R-Users in Wellington:
We are very pleased to advise the presentation of the next meeting of the Wellington R-Users Group. The R-Users Group reflects the community of R-Users within the Wellington area and is about profiling the breadth of the R statistical programming language, assisting users with tips and techniques and supporting the increased use of the language within analytical work. Everybody is very welcome, there are no membership requirements or registration fees, but attendees are encouraged to sign up to the Group on the above website for regular updates.
Refreshments are provided. Please spread the word.
6:00 – 6:05pm Meet & Greet
6:05 – 6:50pm Forecasting the New Zealand General Election with R and Stan
Peter's website features two statistical models that provide probabilistic forecasts for the coming New Zealand election. One is a Bayesian state space model, and the other is based on a set of generalized additive models. Both are implemented in R in combination with the modelling language Stan, and draw on polling data and previous election results from his 'nzelect' R package. In this talk he'll go through the reasoning and assumptions behind the two models and explore a few implementation issues, ranging from optimising code to bring convergence times to a reasonable length, to the logistics of screen-scraping Wikipedia for the latest poll results. This talk is in Peter's personal capacity, and the analysis is not related to any employers past or present.
Peter is a Principal Data Scientist in a public sector agency. Previous roles have included Manager Sector Trends at Ministry of Business, Innovation and Employment, and Director of Program Evaluation for the Australian overseas aid program. In the past six years he has been a leading advocate for the use of R and better analytical practice in the New Zealand public sector. His blog on data and statistics issues using R at https://ellisp.github.io (https://ellisp.github.io/) was recently rated the 13th best statistics blog on the web, and he has several R packages published on CRAN. He is sporadically active on the Cross-Validated and Stack Overflow Q&A sites, and his presence on Twitter - @ellis2013nz - tends to tweet stuff about data and R.
6:50 – 7:00pm Questions / Comments