jakehofman.comhome |
i'm a research scientist in the social dynamics group at yahoo! research new york, where my work involves applications of statistics and machine learning to large-scale social data. i received my ph.d. from columbia university's physics department. please see my resume for more project and background information.
this site serves several purposes, from presenting and organizing my current
research and
teaching
efforts to publishing code and tips
that i hope others will find useful.
i'm on twitter as @jakehofman.
2010.09.28: our paper on predicting consumer activity with web search is released (bbc, ars technica)
2010.05.23: slides and code for my icwsm 2010 tutorial, large-scale social media analysis with hadoop
2009.12.10: our recent paper, "what can search predict?" is posted (blog post, slate article)
2009.10.02: slides for my hadoopworld nyc talk: social network analysis with hadoop
2009.09.11: the course website is up for data-driven modeling (columbia applied math, fall 2009)
2009.05.13: our kdd paper has been covered by mit, slashdot, lifehacker, the chicago tribune, and wired
2009.04.22: our centmail paper was presented and demonstrated at the www 2009 developers track
2008.09.25: we have posted the call for papers for the NIPS 2008 workshop on analyzing graphs
2008.06.24: "a bayesian approach to network modularity" available at physical review letters online
2008.07.04: received "best student presentation award" at mlg 2008 (mining and learning with graphs)
2008.03.18: a scientific american article on my ph.d. advisor, chris wiggins; also a summary from ams
2008.03.14: coverage of analysis of the 2008 aps march meeting co-authorship network from nature blogs
my latest geek tips, also available on twitter, tumblr, or as plain text:
applescript awk bash c c++ consumer css cu cvs debian emacs excel firefox flickr gcc gentoo git gmail google grammar graphviz hadoop html idvd imagemagick iphone iphoto ipod irc itunes java javascript keynote latex lifehack linux mac macosx matlab mobile mysql network networking perl php python quicktime razr rss rstats rsync ruby safari sed sge shell sms sql ssh svn test trac treo unix video windows word wordpress x11
2012.01.19 rstats: use geom_line(aes(group=1)) to connect x-axis factors via a line with ggplot2 (h/t @hadleywickham)
2012.01.18 rstats: a handy syntax reference for specifying statistical models in R: http://bit.ly/yza8SK
2012.01.18 rstats: hide part of a legend in ggplot2 with legend=F in the scale, e.g. scale_size(legend=F)
2012.01.18 rstats: (ab)use ggplot2 to connect x-axis factors via a line with geom_line(aes(x=as.integer(x), y=...))
2011.10.20 macosx: use jot as an alternative to seq for generating sequential and random numbers http://macw.us/pdfadN
my most recent bookmarks at
del.icio.us/jhofman