rugby ready kid

Tuesday October 31, 2006

Filed under: general — jackman @ 1:36 pm

My son, Tom, split his left eyebrow open playing on our bed last night. Splat into the headboard according to Janet (I was at work when it happened): we all met up at PAMF where Tom greeted me with a very cheery “Hi Dad”, which was not bad under the circumstances. I managed to snap the photo below on my phone camera before he was patched up (“oh the memories…”). Tom’s sunny disposition after having his eyebrow split like that suggests he’s “rugby ready”, wahoo… A (pre-treatment!) lollypop, 2 hours and 6 stitches later (not so much fun), we were on our way home.

Janet was the first responder to this one, and held Tom steady throughout the repair job (kudos), and took him back in again today for more sterile tape after he peeled the first lot off overnight.

I looked after JoJo through the whole affair in the waiting area, which has a nice little kids library, and where I introduced her to Alice in Wonderland (a big jump in the words to picture ratio in our family reading).

Comments (6)

post-processing MCMC output

Filed under: statistics — jackman @ 1:07 pm

Richard Zur from the University of Chicago asks:

Dear Dr. Jackman,

… I’m trying to learn more about MCMC post-processing for a project of mine, but I wasn’t able to find the Douglas Rivers article you referenced for the command postProcess [in our pscl package for R]. Could you point me to it and any other post-processing articles you know of?

Thank you,

Good question. Here are some other cites on this, that exploit post-processing of MCMC output, but in slightly different contexts.

Hoff, Peter, Adrian E. Raftery, and Mark S. Handcock. 2002. ‘‘Latent Space Approaches to Social Network Analysis.’’ Journal of the American Statistical Association 97:1090–1098.

Edwards, Yancy D., and Greg M. Allenby. 2003. ‘‘Multivariate Analysis of Multiple Response Data.’’ Journal of Marketing Research 40:321–334.

Comments Off on post-processing MCMC output

Web design is 95% typography

Saturday October 28, 2006

Filed under: computing — jackman @ 1:46 pm

found on daring fireball:

Information Architects Japan » Blog Archive » Web design is 95% typography

Blogged with Flock

Comments (1)

Powerpoint/Keynote: less is more

Filed under: general — jackman @ 6:25 am

found this via digg.com: some people do good Powerpoint, and some people do “insanely great” Keynote….

Comments Off on Powerpoint/Keynote: less is more

animating statistical graphs

Friday October 27, 2006

Filed under: computing,statistics — jackman @ 11:04 am

I occasionally get asked how I create animated statistical graphs, or “statistical movies”. The method I use is to create a long series of sequentially numbered jpegs in R, then string them together in Quicktime. Some recent examples:

  1. traceplot of MCMC iterates, model of weekly levels of approval for George Bush. (126MB Quicktime movie).
  2. imposing normalizing constraints on MCMC output, two-dimensional scaling of the 90th U.S. Senate based on roll call analysis. (66MB mpeg).
  3. 3d fly around a log-likelihood surface (55MB Quicktime movie).
  4. the Law of Large Numbers…! (19MB Quicktime move).
  5. and the ergodic theorem… (20MB Quicktime movie).
Comments Off on animating statistical graphs

first they came for the Vegemite…

Filed under: general — jackman @ 9:58 am

Mo Fiorina brought my attention to a story on Slate about rumors of a ban on Vegemite.

Comments Off on first they came for the Vegemite…

Whitney Music Box

Filed under: computing — jackman @ 9:48 am

Bruce Western brought this way cool site to my attention. Entrancing…

Comments Off on Whitney Music Box

Tennessee Senate Race ads

Wednesday October 25, 2006

Filed under: politics — jackman @ 3:00 pm

I just got told about this ad that aired (is airing!?) in the TN Senate race. Unbelievable… One of the more audacious attack ads I’ve seen. Out here in California it is a pretty sleepy election, and the ads, such as they are, aren’t anything like this. Frankly, Prop 87 is the race for which I am seeing the most advertising, and its tough to get personal and nasty around a ballot proposition.

Comments Off on Tennessee Senate Race ads

The seats-votes curve in 2006; election fraud; election administration

Filed under: politics,statistics — jackman @ 10:18 am

With national polls suggesting that Democrats enjoy a healthy lead as we get close to the 2006 election, a lot of commentary and analysis is starting to look at how the expected national swing towards the Dems will or won’t translate into seats in the Congress. Its a great question: how “flat” or “unresponsive” has the seats-votes curve become, say, through pro-incumbent gerrymandering, strategic decisions by would-be challengers giving big margins to incumbents, etc.

My colleague Josh Cohen directed my attention to a column in the blogs section of the NYTimes, with a simple graph showing how the seats-votes curve has flattened over the post-World War Two period. Nice…

All the polling numbers I’m seeing suggest the swing to the Dems will be bigger than the swing to the Reps when they took the House in 1994, perhaps way bigger. But has the seats-vote curve flattened so much that the Reps could take a big hit on (national-level, aggregate) vote share, but not lose the House? No one I know seems to think so, given that anticipated magnitude of the swing, but lets see.

The blogsphere is also buzzing with rumors of voter fraud ahead of the election: e.g., the Prince of Darkness Karl Rove has a vast network of “fixers” in place to steal the election. I doubt it, I hope not, but above all I wish the United States would get serious about professional, non-partisan election administration instead of this ramshackle tangle of amateur embarrassments and partisan chicanery that passes for elections in this the “greatest democracy on the planet”.

In this connection, I must say that I take special comfort in seeing the Republican candidate for Governor in the State of Ohio, Ken Blackwell, going down big time. This is the same guy who is currently Secretary of State in Ohio and hence is “Ohio’s chief election officer” and “…oversees the elections process and apppints the members of boards of elections in each of Ohio’s 88 counties”. In addition,

He supervises the administration of election laws; approves ballot language; reviews statewide initiative and referendum petitions, chairs the Ohio Ballot Board, which approves ballot language for statewide issues; canvasses votes for all elective state offices and issues; investigates election fraud and irregularities; trains election officials and reimburses counties for poll worker training costs.

So, yes, Blackwell is administering the election in which he is a candidate for governor. And not only that, Ken Blackwell was co-chair of the Ohio Bush-Cheney campaign in 2004, at the same time he administered the conduct of the 2004 election. And of course, Ohio was pivotal in 2004; stories of partisan bias in the way voting equipment was distributed around the state in 2004 are legion, along with the famous (and shortlived) requirement that voter registration applications in Ohio be on 80-pound paper stock. And there are a few stories surfacing already about delays in printing absentee ballots this time around.

In most civilized places, you don’t run elections that way, at least not anymore.

Comments (1)

alpha channel in R graphics

Tuesday October 24, 2006

Filed under: computing,statistics — jackman @ 11:12 am

James S. Coleman Battista writes (inter alia):

So is there another graphing / charting package that will let me overlay one
CI on another with the top one such that I can still see the other one
underneath it? That will let me set a transparency value for the top
one, or that will average or sum the colors / grey values for overlapped
areas without me having to manually calculate the areas to shade them
directly? What I’d really like is something such that a light grey
overlapping a medium gray turns into a dark (=light+medium) grey.

Apologies if this is trivially easy in R. I’ve looked, but I am not
always the best searcher.

see my reply, below the fold.


Comments Off on alpha channel in R graphics
Next Page »

Powered by WordPress