ResearchBuzz

While I was writing the story on the new Merriam-Webster dictionary words I did several experiments with the new words. One of the words I experimented with — locavore — lead me to this great story about a database of over 15,000 growers around the country on a website built by Dan Sutton. His database, LocavoreNetwork, was launched in mid-May and is available at http://www.locavorenetwork.com/.

There’s actually a lot going on on this front page, with information on food safety, some forums (not very active), recommended books and resources for various kinds of gardening, and information on selecting and storing fruits and vegetables. What you’re looking for is the State Information tab, which will take you to http://www.locavorenetwork.com/content/state-information. You’ll still have to use the tab to choose the state whose farms you want to review.

I chose North Carolina. There’s actually two parts to this. First is an “availability guide” that shows you when various fruits and vegetables (from apples to watermelon) are in season. (Hey, where’s the zucchini? Where’s the okra?) The second is the “North Carolina Growers and Producers” section, which is a table (sortable by busines name or county) that show the business name, county, and what they grow, whether it be produce, meat, dairy, manufactured products, or wine (yes, there are wineries listed in this database.)

You can also search the database for various products. Each grower has its own page of details that you can get to by clicking on its name. The details vary a lot; Cane Creek Farm, for example, will give you everything including a full address, phone number, and Web site, while other listings like Green Toe Ground don’t even offer a phone number.

I honestly didn’t think I’d like this site with the way the information is divided by state. But the table is so well-organized and easy to search, and the availability charts so handy (if a bit incomplete) that I’m really impressed. If you’re at all interested in local farming check out this site. If you’re in the US, at least take a quick gander at the availability guide for your state.

This one’s been sitting in my queue for a while; I’m glad I’ve finally got the time to review it. I’m not even sure how long it’s been around. But it’s really good. The Library of Congress has launched Chronicling America: Historic American Newspapers, in beta. The site’s free and available at http://chroniclingamerica.loc.gov/ .

The site currently has keyword-searchable, scanned newspaper pages from 1880 to 1910 in nine states and the District of Columbia. Why so limited? Because it’s still being added to, and more content will be put in over time. There’s also a huge directory of information on newspapers published in the US from 1690 until today. Let’s look at at that first, then we’ll check out the scanned pages.

The front page has a link to the newspaper directory where you can browse by title, but skip that. Go right to the title search page. There you can do the GOOOOOD stuff: narrow down your search by state, county, or city; narrow it by span of time printed, find papers by ethnicity, frequency, or language, and of course search by keyword. And you’ll need to narrow down your search; just running a plain search for newspapers in New York found over 11000 results. *11000*. Yikes.

Results are listed alphabetically and there’s some data available; not enough, but some. Click on a paper name and you’ll get data like geographic coverage, dates of publication, language, frequency, publisher, etc. How much is available varies a lot; once I saw two papers with the same titles whose details varied slightly, but I didn’t get enough information on either one of them to tell them apart. This is really a jumpoff directory; find information on a paper here and use it to move on to searching richer sources.

The search of the newspaper pages, that’s completely different. It’s terrific! The way the search results are displayed is fantastic. But you’ll have to use the search page first: search by state or paper, by year or date range, and then use keywords which can include phrase or proximity search.

Your search results include thumbnails of full newspaper pages! That sounds incredibly unwieldy but the places where your keyword appears are highlighted. When you choose a result you’ll get your page enlarged, again with your keyword highlighted. (I love those highlights — love love love — but I wish they were something besides rose color. Maybe highlighter yellow or nuclear green. If you’re searching for a keyword and it appears only once on a page, you’ll occasionally find yourself in a game of “hunt the highlight.”) You can get the text of the page (though it appears to be machine OCR’d and it looks pretty bad), a PDF, or you can download an image. Best of all, you can use a feature called “Draw Zoom Box,” outline a part of the page you want to enlarge, and immediately you’ll go to that area of the page — with the keyword highlighting intact.

I was amazed at how smooth the zooming transition was. This is the most painless scanned-image newspaper searching I’ve done in a long time. In addition there are so many little extras — getting the pages in a variety of formats, several different levels of paper navigation even at the page-level viewing, and best of all, obvious permalinks to individual result pages. This project is going to be terrific. Between this and Wyoming’s project to digitize its newspapers, I may never read news from this century again. MORE PAPERS!

In an effort to get my information traps back up to snuff, I’m spending some more time messing around with Google News to get the best searches possible.

I keep information traps for both ResearchBuzz material and for my Day Job. For my Day Job, I like to keep track of how universities are using both textbooks and ebooks. While I do monitor the local universities’ newspapers directly, I want to be able to track trends on this usage across the country. This is where Google News comes in really handy.

As you probably know, Google Web search uses the “site:” syntax to restrict your searches to results that come from either one domain (example.com) or a set of “top level” domains (.com). The same syntax works for Google News. So if I want to restrict my news search results to just those that come from university sources, I can add site:edu to my search.

I want to monitor for news about electronic textbooks. My information trap would simply be

e-books site:edu

Of course, I might want to add variants, but that would be my start.

Does site: work with other domains? Sure; for government information you could use site:gov as a search modifier. Want military news? Try site:mil. You can also try those more unusual top-level domains like .biz, .info, and .tv, though it would be harder to narrow those down to one topic or one type of news.

Why not try country codes? I can hear the cool kids in the back saying if those unusual top-level domains should work, then why not use the country codes like .uk, .au, and .ca?

I don’t use country codes when I site search on Google News for two reasons. The first reason is that not all of a country’s media is on a domain that has a country code. For example, I really like The Irish Times. What’s the URL? http://www.irishtimes.com/. Not an .ie to be found. If I did a search for Irish news using site:ie then I would miss The Irish Times. And that would be bad.

The second reason I don’t use country codes with the “site:” syntax is that Google already has a syntax to let you search news by location. That syntax, surprisingly enough, is called “location:”!

“location:”, which is a syntax specific to Google News, allows you to restrict your search results to a particular country (or a particular state — more about that in a minute.) Use the name of the country with the syntax along with any keywords you want to use to restrict your search. For example:

today location:malta

That will find you stories containing the word “today” from the country of Malta. (And though Malta is a small country there are over 500 stories with the keyword “today”.)

I can’t guarantee that all countries will be represented in a Google News search because I haven’t tried all of them. But I found results from countries as small as Monaco:

today location:monaco

So if you’ve got a trap and you want to aim it at a certain country, give that location syntax a try.

And once you’ve tried location: for countries, give it a whirl for US states! Yup, you can use location: with the two-letter USPS postal abbreviation for an American state, and you’ll get news from sources originating in that state. Let’s try this for Rhode Island:

today location:ri

As you can see, that search gets you news from sources like Pawtucket Times, Providence Journal, and Woonsocket Call. You can use this syntax with the 50 states, with the District of Columbia (location:dc) and even with at least one US-affiliated place, Guam (location:gu). (Strangely, Google News did not recognize either Puerto Rico or the US Virgin Islands as valid search locations.)

Knowing that we can search individual state news with the location: syntax, let’s go back to the site: syntax. Can you combine site: and location: syntax? Absolutely! Say I want to find college and university news in North Carolina. Nothing easier:

site:edu location:nc

You don’t even have to use any search terms with that, though you will get a LOT of results. If I wanted to restrict my search for textbook or ebook information to one state, this would be the way to do it. Or if I wanted to focus my traps on one topic for one state, I could do it this way too:

basketball site:edu location:ca