Computing

Software I wrote

  • RRHO R package: Implements the Rank-Rank Hypergeometric Overlap Test. Download from Bioconductor or Github.

  • chords R Package: Estimate population size and degree distribution in respondent driven samples, presented in our paper. Download from CRAN

  • SimpleLearner: Implements an easy to tune polynomial deep learning network. Download from GitHub.

  • Scraping ArXiv: Script for scraping ArXiv meta data and doing some exploratory statistics of manuscript ArXiving. Download from github.

  • SelectiveCI R package: Implements post-selection confidence intervals, presented in our paper. Download from github.

  • FPF R Package: Finds the prevalence of activation in fMRI group studies, presented in our paper.

Software I use

  • Ubuntu- a friendly Linux distribution with Hebrew support.
  • Data:
  • Typesetting:
  • Version control: git with github for code, manuscripts, etc.
  • Scribus- for scientific posters.
  • Zotero- for managing references (I like Mendeley, but I am already addicted).
  • Calibre for managing my (e)book library.
  • Reeder- by far my favorite RSS reader (since GoogleReader closed).
  • Pocket- For all the “I want to read this later” sites.
  • Jekyll with GitHub: for blogging.
  • Linux Shell and system administration:
    • zsh: I just like it better than bash.
    • fish shell. Great shell (but don’t make it default as it is not POSIX compatible).
    • Filezilla: a friendly GUI for your file transports.
    • SSH Tunnel Manager: A friendly GUI for ssh tunneling.

Software I want to try but didn’t have the time yet:

Hardware I use

  • Server- Lenovo X3750-M4: Just because I find that the batch cluster computing paradigm (e.g. Condor, SGE, etc.) is not suited for modern interactive data analysis. I found that you can get much more done with a single, massive, multiple core, RAM rich, server (running RStudio Server obviously).
  • Laptop- Lenovo T460s: The size of the X1 with the power of the T series.
  • Cloud- AWS: For any ad-hoc hardware needs (which is quite often for me), I just start an EC2 instance.