ResearchBuzz

Google’s hard at it in the labs again with a new feature called Living Stories. Living stories are more like living “topics,” with story topics put together into a permanent place and updated in a variety of ways. It’s available at http://livingstories.googlelabs.com/.

Once you get to the site, pick a story you want to follow. I chose executive pay, which is at http://livingstories.googlelabs.com/lsps/executivepay. Note the permanent URL. When you get there you’ll see a page which looks like this:

There’s a summary at the top of the page, a timeline, and down the middle of the page stories. The left side of the page beneath the timeline provides pointers to different kinds of media/information, while the right side provides links to important stories in the timeline.

When you click on a story you find interesting, you don’t leave the page — it opens right in the middle, which does make it easier to take in several stories without losing your place or getting hugely distracted. Key people and companies in the stories are highlighted; click on them and you’ll get a popout with a one-sentence explanation and sometimes a picture.

A took a look at the supplemental materials. There were plenty of images and quotes available, fewer videos. There were some great graphics showing the evolution of executive pay.

So how do you keep up with the changes to the stories if the URL is permanent? You have a couple of options. The first is a good old-fashioned RSS feed. The second is signing in with a Google account and getting updates to the story e-mailed to you.
Personally I’d prefer RSS.

Google’s put together a great way to group a lot of stories in one place. The one problem is the sources. The New York Times and Washington Post are great papers, but there are lots of other great papers, too, and you could get an even more multifaceted look at
a story if you used several different sources. I’ll be a lot more interested in this when I can go to a story about something happening in Virginia, for example, and get stories from every indexed news source in the state.

I love the RSS feed for New York Times Open. The blog doesn’t update very often but whenever it does I know I’m going to have something good to read. At the end of October the blog announced the release of 5,000 of its person name subject headings as “Linked Open Data.”

Let’s back up. The New York Times has developed subject headings to index its archives. So you can think “tags” instead of “subject headings” if you like. (Sorry NYT.) And the “Linked Open Data” means that the 5,000 subject headings/tags have been manually mapped to the data sources Freebase and DBPedia. Two very exciting things about this: 5,000 is just a fraction of the NYT’s subject headings — there are in total over 30,000 — and the NYT intends to map and release them all. Further, the NYT is releasing this data under a Creative Commons license!

You can explore what’s been done so far at http://data.nytimes.com/. You can download all the data records in one file (you have to agree to the CC license first) or you can browse by last name. I went and looked at the E’s to see if Elmo had been indexed. He hadn’t, but there were several other people under E, from Eagleburger, Lawrence S, to Eyre, Richard. Each name has an URL associated with it. Click on it to get more data.

I clicked on Herm Edwards’ name which is located at http://data.nytimes.com/57985207950391437243.html. Data here includes the number of mentions in the NYT, first and last time the subject heading was mentioned (I’m suspicious of that “first time” — it seemed like 2001 in most of the names I looked at and that didn’t seem right for the historical figures) and a pointer to the New York Times “Topic Page”. (If you’re looking for the latest news and other information on a figure, in an easy-to-read format, use the Topic Page. It even has an RSS feed.)

Even more interesting than the aggregated data are the pointers to Freebase and DBPedia. Each person’s data URL is also associated with links to pages of data at Freebase and DBPedia. These two pages are in XML and RDF (Resource Description Framework) formats respectively, so they’re less for reading by humans and more for mixing and reoutputting by computer programs.

The NYT release of these subject headings/tags helps pull three data sources together. I expect to see some great tools made from this. If you’d like to see how people are discussing extending and using the new release, you can check out the Linked Open Data community at http://groups.google.com/group/nyt_linked_open_data.

Hat tip to ReadWriteWeb for letting me know about the New York Times’ recent Custom Feeds tool, available at http://prototype.nytimes.com/customFeeds/.

Essentially this tool gives you the ability to create keyword-based RSS feeds for New York Times content, but also enhances that essential idea with a couple of extras.

Here’s a snapshot of the site. First thing you do is enter a topic or keyword in which you’re interested. The tool will suggest other terms based on the keyword you initially entered. (As you can see I entered Washington and the NYT had lots of suggestions for that.) Next you choose which keywords you want to add to your custom feed. (You can add several if you like.) When you enter keywords, the tool will evaluate the feed, testing to see how many articles in the past 30 days include your term(s) (or, if it’s a very active term, how many articles in the last one day included your term(s).) This is AWESOME; you’ll know right away if a feed is going to contain too much/too little content without having to run tests.

Once you have a term or set of terms you like, enter a title for the feed and click the “Subscribe” button. The NYT will kick out a RSS file that has a pretty good snippets from NYT articles as well as the occasional image.

The only thing that even bothered me a little bit about this tool is the fact that you have to enter your own feed titles, which would slow things down if you wanted to create a lot of feeds. Other than that these extras as terrific. Highly recommend this tool.