Software

I am co-author of the selectiveInference package on CRAN. Additional code is on github.

Methodology

I am interested in high-dimensional statistics and Gaussian processes. Much of my research has been focused in a new area called selective inference. The approach combines model selection and inference by conditioning on the model selection event and performing inference via the associated truncated distribution. Often the resulting significance tests are slightly modified versions of the classical tests in regression analysis. As an easier introduction to this work than the papers referenced below see the slides from my invited presentation at JSM2015.

J. R. Loftus. Selective inference after cross-validation. Preprint.

J. R. Loftus and J. E. Taylor. Selective inference in regression models with groups of variables. Preprint.

J. E. Taylor, J. R. Loftus, and R. J. Tibshirani. Tests in adaptive regression via the Kac-Rice formula. To appear in the Annals of Statistics.

J. R. Loftus, J. E. Taylor. A significance test for forward stepwise model selection. Preprint.

X. Tian, J. R. Loftus, J. E. Taylor. Selective inference with unknown variance via the square-root LASSO. Submitted.

Works in progress

J. R. Loftus and M. Baiocchi. Valid inference for the largest treatment effects among heterogeneous subgroups. Work in progress.

L. Blier, J. R. Loftus, J. E. Taylor. Inference on the number of clusters in k-means clustering. Work in progress.

Applications and collaborations

As a trainee in the new Biostatistics for Personalized Medicine program, I have coursework and collaboration related to genomics and analysis of data from new sequencing technologies. I have also worked on a variety of projects with data from diverse areas during my time at Stanford.

Developed a permutation testing procedure to account for non-independence in reads from ATAC-seq in collaboration with the Chang lab.
One of three students recruited into the inaugural session of the Stanford DataLab, learning data science best practices while working on a project to alleviate poverty in Kenya.
E. T. Richardson, et. al. Modeling cash incentives vs. oral preexposure prophylaxis in high-risk African women: the Cash-PreP Study. Poster, AIDS Conference.
Team leader at 2013 DataFest competition. We won a $1,000 prize for our project applying causal inference to US Senate electoral finance data.
Conducted extensive crowd-sourcing experiments during an internship at Google in 2013.
Frequently participated in and more recently organized the consulting service offered by the Stanford Statistics Department.