Thursday, December 30, 2010
At the end of the year I got to thinking about life in LA and all the oil shocks and crises that have been in the news during 2010. And that made me think of a simple enough question:
When was the first productive oil well drilled in California? And.. for extra credit, who drilled it and where?
I'll warn you--this seems really simple, but answering the question definitively took me a while (and I'm still triple-checking my answer).
For clarity, a "productive" oil well is one that pumped more than 10 barrels of oil AND the output of which was refined and not used in its native out-of-the-ground state.
Tuesday, December 28, 2010
As we've seen over the past year, Google (and Bing) have made many changes to their interfaces, user-facing capabilities, underlying index and results ranking. The most visible changes were to Bing's image search interface (continuously scrolling results) and Google's Instant Search ("search results come back as quickly as you type"). But many, many changes get made every week--some obvious, some pretty subtle.
The biggest news story in many ways was the "DecorMyEyes" story which broke in the New York Times in late November. Good investigative journalism on the part of the Times revealed a fairly abusive reseller who was taking advantage of his obnoxious behavior (which generated a large number of web posts linking to his site) to boost his position in searches for his product.
But more interestingly (for us on SearchResearch), Google fixed this problem within just a few days by a clever ranking algorithm tweak. As described by Amit Singhal on the Official Google Blog, the solution is more than just sentiment analysis, but involves detecting overall terrible user experiences on the part of purchasers and then using that information to change the rank position of results of sellers.
The implication of all these changes--which are ongoing and continuous--is that web search is a dynamic beast. What you get from a search today might not be what you get tomorrow.
In other words, web search isn't anything like a normal "reference search" from the days of yore. Not so long ago, reference materials stayed pretty constant, or at least changed slowly enough that the book / journal publishing cycle was rapid enough to stay up with the changes.
Now, however, things change rapidly. Not only does information accelerate (a point made masterfully by James Glecik in his book Faster: the acceleration of just about everything), but aggregations of information are constantly bubbling, as are the tools by which you access the information stew.
Point: Stay in touch with the changes going on. For the most part things-will-just-work. But when they stop working, you'll want to know how and why, especially if you're trying to make sense of a complex world.
Saturday, December 25, 2010
1. Google search cheat sheet–there are many Google cheat sheets out there, and this is mine. This one has the benefit of actually being correct. It's also available as a mousepad, if you'd like to have one of your own. It shows about 20 of the top tricks and search operators that are most useful. Print it out and distribute widely. You have my permission.
2. Reading level search–If you somehow missed the announcement, Google recently launched the ability to filter search results by reading level.
3. Creative Commons license search–Google also recently launched another advanced feature in Image search. When searching for images, you can also go into the Advanced Search mode and filter by CC license level.
4. Custom Search Engines–A CSE lets you create your own "mini-Google" that searches just over the sites you like. That means it's really easy to create a special-purpose search-engine for just the needs of your class... or even a specific lesson. I'll do a posting about this in the future, but if you want to get started exploring, click on the link above. It's actually very easy to do and solves all kinds of problem when letting younger searchers look for specific topics on the web.
5. Alerts–A Google Alert is a standing query that's automatically run for you on a daily or weekly basis. Any changes in the web search results (or News) are automatically sent to you as an email. Think of the Alerts tool as your personal assistant who is always scanning the net for you. (I'll also write a longer post about this as well.) I have Alerts set up for my name (so I can see who's talking about me!) and for four different topics I'm interested in. Naturally, one of those topics is "how to teach search skills," which I have set up to send me weekly updates. It's a very handy way to track the latest in your special topic of interest.
Hope you like these. More to come in the days ahead.
Have a wonderful holiday!
Searching on into 2011...
Thursday, December 23, 2010
The best laid schemes o' mice an' men gang aft a-gley.--Robert Burns ("To a Mouse"). Fred and JPP also found this out, and give the entire stanza as:
In proving foresight may be vain:
The best laid schemes o' Mice an' Men,
Gang aft agley,
An' lea'e us nought but grief an' pain,
For promis'd joy!
Wednesday, December 22, 2010
Friday, December 17, 2010
Which is why you really want to listen to someone saying the name, rather than trying to figure it out from a pronunciation guide. I don't know about you, but I can't imagine what "L" + "th" + "hiss" should really sound like.
My next query was intended to look for the category of names (rather than just the single name Llewellyn) and also look for pages that would have recordings on them.
What's the most common way of describing pages with sound bites? I'm willing to bet they ALL say something like "hear these names spoken aloud" (or equivalent language). So I chose to use the term 'hear' and the category 'Welsh names' as in this query:
In yesterday's comments, Hans points to the http://www.pronouncenames.com/pronounce/llewellyn page, which has a nice "standard English" vs. "Welsh" pronunciation side-by-side. As Hans' solution points out, there are multiple ways to find a good answer to the original question.
Thursday, December 16, 2010
Saturday, December 11, 2010
This really WAS a hard search problem! Congrats to Hans and Fred for figuring it out.
Here are my search steps:
* I found in Google Maps the name of a bay in the neighborhood "Brosewere Bay"
* A search in Google with ["long island" "brosewere bay" +canals] brought me to a website of the Long Island South Shore Estuary Reserve:
[ long island wetland ditches ]
* A search on that site using Google [site:www.estuary.cog.ny.us ditches] brought me to a document mentioning grid ditching for mosquito control http://www.estuary.cog.ny.us/ISR2005/ISR%20Outcome%204.pdf
* A search in Google on ["long island" grid mosquito control] came up with the final document: http://www.geo.sunysb.edu/lig/Conferences/abstracts07/abstracts/potente.pdf
Friday, December 10, 2010
Wednesday, December 8, 2010
The question is just what I wondered as we flew overhead. What's the story with those straight channel-like links? Can you figure out when those channels were created and what they were trying to do?
Tuesday, December 7, 2010
Example: To find salsa recipes that do NOT have tomatoes in them,
[ salsa recipe -tomatoes ]
That does just what you think, returning results that do NOT have the word "tomatoes" on the page.
So.... what does + do?
Most people think it means the opposite of minus, but, alas, is doesn't. (Although I agree that would make sense!)
This what + means: DO NOT change the search term in any way. No synonyms, no stems, no nothing. Search for exactly this word. (In other words, it's very much like double-quote for a single word.)
Example: My Mom told me yesterday that she was going to hear some "authentic joiker music" at the local Scandanavian outlet. She was excited about it, and I naturally asked "so... what's joiker?"
A quick Google search for joiker is pretty useless--all of the results are about JOKER as the word gets spell-corrected to something I didn't really want.
This is when you want to use +
[ +joiker ]
And that tells you pretty quickly that it's a traditional Sami (aka Lapplander) style of singing. It's pretty interesting, actually... Wikipedia tells us (with a spell-correction to yoik) that a joik is a song that tries to "transfer the essence" of a person or place to the listener, rather than being "about" a person or place.
In other words, + is the same as the Bing command noalter: (example, on Bing: [ noalter:joiker ] )
Double quotes (on both Google and Bing) serve to turn off synonymization for strings of words. Example:
[ "joik music" ]
Thursday, December 2, 2010
Unfortunately, that page only goes back as far as 1998--but it's a pretty good solution!
If we want to go back further in time, the only way I know to get this data is through the Rand Corporation's. I found this factoid out by doing the query:
[ california annual per-pupil spending data set ] - since I was looking for the complete data, it was pretty easy to see that only Rand had all the data compiled together.
BUT... once you click through to Rand's data set page http://ca.rand.org/stats/education/perpupil.html -- you quickly discover that you need to login to actually get the data. Unfortunately, logins start at $270 / person.
Luckily, I happen to know that the Rand data sets are often made available through public libraries! (Keep this in mind: this is true for MANY data resources that otherwise cost a lot of money.)
So I connect to my local library's page (in my case, the Palo Alto library lists all of its online resources on a web page, which makes it easy to find).
All I did was then click through to the Rand database site for California Education Statistics, login via the interstitial page (requires only my Palo Alto library card), and voila! -- I'm into the Rand dataset.
Once there, it's an easy navigate (Databases>Annual spending per pupil) to the list of counties of California, and then a quick download of the data from 2000-2009 as a TSV file (easy to then import into your favorite spreadsheet program).
Here's an example chart from the data. You can easily see the bug in the data (no, the Montebello school district did NOT spend $47K / student).
The moral of this particular story?
Sometimes you still need to find out where the data is kept... and that the access path might be through your public library!
Wednesday, December 1, 2010
Wednesday Search Challenge (Dec 1, 2010): Finding data you can use to compare school district per-pupil spend
And, you'd like to see the historic trends since 1990.
Challenge: You're looking for an assembly of data for all the California school districts, per-student spend, per year. Ideally, you'd like a graph that lets you compare your school district with all the others in the state.
Yes, I know you can buy such data. But your challenge is to do this little research task on a typical school budget.. which is to say, for free.
Start your search engines!
Tuesday, November 30, 2010
Thursday, November 25, 2010
The big search lesson to take from the answer to this challenge is this: Don't assume too much about what you think the answer will be. Your initial assumptions might be very wrong, and you'd waste a lot of time trying to prove something that just isn't true.
Here's the story...
But how do we search for this kind of information?
My first search was [ cranberries 1621 ] which gives a number of search results, the most interesting of which is the "The Truth About Thanksgiving" from the Planet Blacksburg (VA) news site that quotes Daniel Thorp (colonial history prof at Virginia Tech)
After looking through lots of deadend links (research is a slow process!) I finally decided to look for the original letters describing the Thanksgiving holiday in 1621 to see what I could find. From the Daniel Thorp article I found that the letter was written by Edward Winslow, making the obvious search
[ Edward Winslow 1621 ] lead me directly to a transcription (and handy translation into modern speech): http://www.pilgrimhall.org/1stthnks.htm
I quote from their site (in modern spelling):