This blog post regarding a reanalysis of a mobile phone app deployment and its related paper, “Multi-list Mark-Recapture methods: A semi-mathematical tutorial with illustrations”, is now available.
Causal inference for prediction of streamflows and other phenomena (tenative title, in progress)
This short-term project and activity has been started.
A committee of the Town of Westwood, its Environmental Action Committee, sought to baseline shopping bag use prior to launching a campaign to promote bringing own bags to stores. They asked me to design a sampling plan which used the modest volunteer resources of the Committee to ascertain bag use rates by type. A follow-on study is begin planned to judge campaign effectiveness.
An illustration of the use of penalized B-splines to produce best smooths for stream flows of Mother Brook in Dedham. The best smooths are picked using generalized cross-validation.
A question was posed by a colleague a couple of months ago: What are the odds of a stock closing at the same price it opened? I found the question interesting, because, at first, it appeared to be a one-dimensional version of another problem I looked at, which was in two dimensions. Well, I have produced an estimate, and am reporting results here. My first impressions of the problem were wrong. It actually is a two dimensional problem, not a one dimensional one. And it is not the same as the earlier problem, because although one of its dimensions is discrete, the other, time, is (essentially) continuous. I explain at the post.
The HADCRUT4 and GISTEMP series are two prominent time series for global temperature. This study examines use of dynamic linear models for smoothing these series, estimating their uncertainties in the smooth, and then calculating a Bayesian combined smooth using both series together.