News for hackers. Stuff that matters.
Re:Google Labs: Books Ngram Viewer by jambo on Thursday December 16, 2010 @03:38PM (#2013550)
I'm very happy to see Google offering the ngram counts for free download (http://ngrams.googlelabs.com/datasets), which I wasn't expecting, given their earlier web corpus data release on DVDs. Maybe we'll see an update on that front, too.
Re:Google Labs: Books Ngram Viewer by boredguy8 on Thursday December 16, 2010 @06:54PM (#2014434)
The "long s" causes problems with OCR and may be an easy-ish fix, if someone from Google were to take a look.

http://ngrams.googlelabs.com/graph?content=fame,same&yea...

http://en.wikipedia.org/wiki/Long_s

Re:Google Labs: Books Ngram Viewer by nl on Thursday December 16, 2010 @06:17PM (#2014316)
Most interesting I've seen so far:

Wall Street vs Silicon Valley: http://ngrams.googlelabs.com/graph?content=wall+street,silic...

Telephone, Internet: http://ngrams.googlelabs.com/graph?content=telephone,+intern...

Microsoft, Netscape, Yahoo, Google, Facebook: http://ngrams.googlelabs.com/graph?content=Microsoft,Netscap...

Great War, Hitler, Communist, September+11: http://ngrams.googlelabs.com/graph?content=Great+War,Hitler,...

Re:Google Labs: Books Ngram Viewer by zacharypinter on Thursday December 16, 2010 @09:28PM (#2014878)
Another fun one:

terrorism,socialism,communism: http://ngrams.googlelabs.com/graph?content=terrorism%2Csocia...

Re:Google Labs: Books Ngram Viewer by ovi256 on Thursday December 16, 2010 @06:45PM (#2014410)
What are those weird cycles in the Wall Street data ? Is that in any way related to the bubble/retraction cycle ?
Re:Google Labs: Books Ngram Viewer by nl on Thursday December 16, 2010 @08:17PM (#2014665)
Here's US GDP growth: http://www.google.com/publicdata?ds=wb-wdi&met=ny_gdp_mk...

This could be a better chart: http://en.wikipedia.org/wiki/File:GDP_growth_1923-2009.jpg

There might be some kind of relationship there, but I don't think it's the whole explanation.

Re:Google Labs: Books Ngram Viewer by jjclarkson on Thursday December 16, 2010 @06:16PM (#2014314)
I think overuse of the word "cool" might finally be heading downward. And it's interesting to see the blips in the 70's and 80's of "groovy" and "rad" respectively.

http://ngrams.googlelabs.com/graph?content=cool%2Chip%2Cgroo...

Re:Google Labs: Books Ngram Viewer by akozak on Thursday December 16, 2010 @03:58PM (#2013645)
This is really fun to play with: http://goo.gl/mNSDD
Re:Google Labs: Books Ngram Viewer by shib71 on Thursday December 16, 2010 @04:08PM (#2013687)
Why would you use a shortener when there is no need? Especially when the URL is useful as it is here.
Re:Google Labs: Books Ngram Viewer by akozak on Thursday December 16, 2010 @05:13PM (#2014006)
It's fun to track how many clicks it gets, and I'd already had it on my clipboard for twitter :)
Re:Google Labs: Books Ngram Viewer by pkamb on Thursday December 16, 2010 @04:33PM (#2013797)
I think it was an atheist rick-roll.
Re:Google Labs: Books Ngram Viewer by makmanalp on Thursday December 16, 2010 @04:35PM (#2013811)
Re:Google Labs: Books Ngram Viewer by akozak on Thursday December 16, 2010 @05:15PM (#2014019)
It's actually case-sensitive, so here's a more accurate comparison: http://ngrams.googlelabs.com/graph?content=Christianity,Isla...
Re:Google Labs: Books Ngram Viewer by msbmsb on Thursday December 16, 2010 @06:56PM (#2014438)
Country names/adjectives: English,England,British,French,France,German,Germany

In most cases the adjectival form of a country tracks pretty closely to the nominal form (e.g. France vs French and others) except for the case of England vs English/British, the latter two spike from almost nothing around 1795. I can't think of a good explanation for why those two forms were so rare in comparison with "England" before 1795, anyone have an idea?

http://ngrams.googlelabs.com/graph?content=English%2CEngland...

Re:Google Labs: Books Ngram Viewer by msbmsb on Thursday December 16, 2010 @04:46PM (#2013876)
Re:Google Labs: Books Ngram Viewer by srean on Thursday December 16, 2010 @05:42PM (#2014160)
The first thing I tried out was different 3-grams of "Colorless green ideas sleep furiously". Here are the results

http://ngrams.googlelabs.com/graph?content=Colorless+green+i...

The correlation patterns and the year where the curve picks up are interesting. It would be illuminating to map the peaks.

Re:Google Labs: Books Ngram Viewer by goldins on Thursday December 16, 2010 @02:56PM (#2013354)
An interesting thing I noticed is that the usage of the word "the" has decreased by about 1% in the last century (which I think is significant considering the data). I wonder what the reason for this is.
Re:Google Labs: Books Ngram Viewer by yoak on Friday December 17, 2010 @02:51PM (#2017466)
Using American English throughout:

inflammable, flammable: http://ngrams.googlelabs.com/graph?content=inflammable%2Cfla...

regardless, irregardless: http://ngrams.googlelabs.com/graph?content=inflammable%2Cfla...

Re:Google Labs: Books Ngram Viewer by nl on Thursday December 16, 2010 @06:38PM (#2014389)
I love this. Here's the story of the world since 1750. The scientific revolution, followed by the industrial revolution, followed by the computer revolution, with the internet revolution just starting:

http://ngrams.googlelabs.com/graph?content=science,industria...

Re:Google Labs: Books Ngram Viewer by igravious on Thursday December 16, 2010 @05:11PM (#2014000)
Hey Google!

Combine this with Google Sets ( from Labs as well http://labs.google.com/sets?hl=en&q1=immigration&q2=... ) so that I can feed a list generated by Sets into Ngrams.

Thanks! (And a pony too.)

Re:Google Labs: Books Ngram Viewer by wgrover on Thursday December 16, 2010 @03:57PM (#2013643)
Try "war": (http://ngrams.googlelabs.com/graph?content=war&year_star...). Nice peaks on WWI, WWII, and Vietnam; maybe a tiny peak on the Gulf War. Any other search terms where you can see meaningful fluctuations?
Re:Google Labs: Books Ngram Viewer by rbancroft on Thursday December 16, 2010 @04:39PM (#2013841)
Yes, that was a cool one.

murder: pretty cyclical and consistent rape: massive spike in the 80s and 90's, and declining a bit

fraud: blip after WW2, then pretty stable up to around 1995

nazi: interesting trends there

oil: some more interesting trends, a large spike in the early 80's

Interesting that the trends for freedom and liberty differ quite significantly, although they have some of the same 'humps'. Liberty has an uptick after 2000, whereas freedom does not.

Re:Google Labs: Books Ngram Viewer by igravious on Thursday December 16, 2010 @05:05PM (#2013965)
And you didn't try War and Peace? I did. Which term occurs more often in the corpus do you reckon? Try going back to 1500. Are you surprised? How about trying Love and War for all you hippies out there. What does this all say about humanity?
Re:Google Labs: Books Ngram Viewer by jluxenberg on Thursday December 16, 2010 @05:34PM (#2014118)
Re:Google Labs: Books Ngram Viewer by tudorw on Thursday December 16, 2010 @04:09PM (#2013693)
Re:Google Labs: Books Ngram Viewer by tudorw on Thursday December 16, 2010 @04:12PM (#2013705)
Re:Google Labs: Books Ngram Viewer by staticshock on Thursday December 16, 2010 @04:12PM (#2013709)
i'm surprised that "star wars" has never been able to eclipse "galoshes":

http://ngrams.googlelabs.com/graph?content=star+wars,+galosh...

Re:Google Labs: Books Ngram Viewer by hanibash on Thursday December 16, 2010 @06:05PM (#2014261)
Someone please explain why "I love you" is cyclical???

http://ngrams.googlelabs.com/graph?content=i+love+you&ye...

Re:Google Labs: Books Ngram Viewer by nl on Thursday December 16, 2010 @06:22PM (#2014337)
If you remove the smoothing it looks more spiky then cyclical: http://ngrams.googlelabs.com/graph?content=i+love+you&ye...

It's possible that's more to do with the distribution of books Google has scanned than anything.

Re:Google Labs: Books Ngram Viewer by nl on Thursday December 16, 2010 @08:04PM (#2014628)
WTF? Why the downvote!?
Re:Google Labs: Books Ngram Viewer by jpwagner on Thursday December 16, 2010 @06:25PM (#2014347)
wars and birthrates
Re:Google Labs: Books Ngram Viewer by jpwagner on Thursday December 16, 2010 @06:04PM (#2014258)
Re:Google Labs: Books Ngram Viewer by jpwagner on Thursday December 16, 2010 @06:05PM (#2014263)
Re:Google Labs: Books Ngram Viewer by ddlatham on Thursday December 16, 2010 @06:11PM (#2014290)
Re:Google Labs: Books Ngram Viewer by boredguy8 on Thursday December 16, 2010 @05:51PM (#2014196)
quick brown fox, brown fox, slow brown fox

http://ngrams.googlelabs.com/graph?content=quick+brown+fox,b...

Am I reading that correctly: People rarely talk about brown foxes without mentioning their speed, which is never slow?

More importantly: if I wrote and published a book about a slow brown fox, could I potentially do to foxes what Chomsky did to green ideas?

Re:Google Labs: Books Ngram Viewer by boredguy8 on Thursday December 16, 2010 @06:15PM (#2014305)
Re:Google Labs: Books Ngram Viewer by gmartres on Thursday December 16, 2010 @03:52PM (#2013621)
Great! It just lacks an option for case-insensitive search.
Re:Google Labs: Books Ngram Viewer by msbmsb on Thursday December 16, 2010 @04:25PM (#2013765)
one,two,three,four,five,six,seven,eight,nine,ten

http://ngrams.googlelabs.com/graph?content=one%2Ctwo%2Cthree...

In order, except "ten" shows up between "six" and "seven".

Re:Google Labs: Books Ngram Viewer by igravious on Thursday December 16, 2010 @04:59PM (#2013934)
That's interesting, isn't it? I think this proportionality is related to a quirk of human perception. If you are shown briefly a set of objects the average person is able (at a glance) to quickly count those objects for a low value whose threshold is around six or seven. Thus when the number goes any higher the individual will respond that there were "less than ten" or "fewer than ten". Of course because humans have standardized on base 10 as there counting system then ten will get bumped up in any distribution but to predict that it would occur between six and seven (the number of man and divinity no less) would be difficult.

What about 11 to 20? Hmm, it (roughly) goes: twenty, twelve, fifteen, eleven, eighteen, fourteen, sixteen, thirteen, seventeen, nineteen! Just what I expected :)

Re:Google Labs: Books Ngram Viewer by dwiel on Thursday December 16, 2010 @05:54PM (#2014210)
1,2,3,one,two,three

http://ngrams.googlelabs.com/graph?content=1%2C2%2C3%2Cone%2...

for a short period of time in the 80s and 90s, 1 was more popular than one, where from the 1700s through 1950s, 1 was half as popular as one ...

Re:Google Labs: Books Ngram Viewer by msbmsb on Thursday December 16, 2010 @05:16PM (#2014027)
Also interesting, the 1700s were not kind to the numbers "six" or "seven":

http://ngrams.googlelabs.com/graph?content=one%2Ctwo%2Cthree...

Re:Google Labs: Books Ngram Viewer by ivancho on Thursday December 16, 2010 @05:07PM (#2013973)
Re:Google Labs: Books Ngram Viewer by msbmsb on Thursday December 16, 2010 @06:46PM (#2014414)
Exactly
Re:Google Labs: Books Ngram Viewer by nl on Thursday December 16, 2010 @05:47PM (#2014177)
I love the "whale fail" image you get if you have an error: http://imgur.com/qCDCj
Re:Google Labs: Books Ngram Viewer by dchest on Thursday December 16, 2010 @04:11PM (#2013702)
"computer" and "internet" reveal Y2K problem.
Re:Google Labs: Books Ngram Viewer by programd on Thursday December 16, 2010 @03:56PM (#2013636)
There has been a ginormous increase in the use of the word "ginormous" in the last decade.
Re:Google Labs: Books Ngram Viewer by ntoshev on Thursday December 16, 2010 @05:32PM (#2014101)
Google trends for books?
Re:Google Labs: Books Ngram Viewer by Jun8 on Thursday December 16, 2010 @04:24PM (#2013763)
Try "nerd". Like what you would expect. Now, try "sexy nerd". Sigh!