davemessina.com » science

Data come and me wanna go home

Dave — Tue, 13 Nov 2007 09:24:08 +0000

Today my good friend Todd pointed me to Swivel, apparently being touted as “YouTube for data”.

Seems promising, although most of the datasets are vanishingly small, and too “munged” — not raw enough for reanalysis. For instance, there seems to be a lot of stuff like this:

Billion Dollar Climate and Weather Diasasters, 1980-2006

which comes straight off a website and is already heavily processed. To their credit, you can download their technical paper on this analysis, but it’s scant on methodology and the source data is unavailable.

I guess there’s value in trying to centralize all of these bits of data, analyzed or not, and the already-analyzed stuff is of course far more interesting:

average trans fat in grams by fast food restaurant

But still: where’s the everyone-can-edit part, like Wikipedia? It’s one thing to be able to leave a comment, but the strength of this site could be the ability for those knowledgeable in data analysis, a subject area, or both to come in, review, and actually correct or reinterpret the data. Letting people simply throw up whatever crap they happen to believe could make the problem of data-and-statistics misuse even worse, because there will be no way to evaluate the findings.

I can’t wait to see what Edward Tufte has to say about this.

Update 13 nov 11:46a GMT+1:
I checked out the Swivel blog and (unsurprisingly) it gave me a better idea of the creators’ vision. And some cool links, too, to other data-related sites.

[image courtesy of twid]

The art of science

Dave — Thu, 17 May 2007 03:23:00 +0000

My friend Victoria pointed me to a cool website of vintage science and science fiction prints, Droppin Science. They even have some of those hokey filmstrips they used to show us in grade school.

A better way to manage journal articles

Dave — Thu, 26 Apr 2007 04:50:54 +0000

Sure, there’s Endnote. But it munges your manuscripts. And generally behaves like an app whose marketshare has long exceeded its quality. Not to mention — $239.95?

Enter Papers. Developed by Mekentosj, two biologists with quite a track record, Papers has the creativity and attention to detail that make me want to play with it, adding PDFs and Word docs to it that I don’t really need to keep track of.

While there are still a few interface bugs and crashes, it is a 1.0 app. What it really needs is the ability to cite your stored articles and reformat them according to a journal’s requirements. The authors say that capability is already well-covered by Endnote and its competitors. But heck, Papers has much of those programs’ other functionality already.

Still, you should check it out, if only to see the addictive swash-swash effect when you grab an article’s metadata from Pubmed.

April 26, 2007: Papers v1.1 came out today, and so far seems to have fixed the most serious bugs.