r

Background One of the challenges of developing clinical prediction models is deciding which one is best. Sometimes traditional performance measures such as the Brier score or C-statistic might be used to quantify to what extent one model performs better than another.

Background Bootstrapping is a helpful technique for identifying the variance of an estimate in a given sample when no other data are available. In the case of the evaluation of clinical prediction models, bootstrapping can be used to estimate confidence intervals around performance metrics such as the Brier score, the c-statistic, and others.

Every now and then it is useful to make a map. In times of political uncertainty, data can light a path forward. Your local elected officals may be interested in data, too, and how they impact policy (and re-election, of course).

Here’s the blog post originally posted on babelgraph.org on April 4, 2011. I suspect the twitter API and much of the code below are not up to date. I am a recent comer to twitter, and it took me a few weeks to figure out what this was all about.

Not infrequently I am in a research meeting and someone says, “It would be really cool to get data on travel times for people. But I don’t know where to find that.

Here’s the blog post originally posted on babelgraph.org on April 14, 2012. In a recent blog post by CMastication, a little meme puzzle is presented with the introduction that a preschooler could solve it in 5-10 minutes, a programmer in an hour.

Here’s the blog post originally posted on babelgraph.org on July 11, 2012. Thanks to Hadley Wickham for referencing some of content here, and apologies for the broken URL. NB. The original C++ code didn’t seem to compile on my computer today.