Google Labs: Books Ngram Viewer

http://ngrams.googlelabs.com/

by abraham • 16 years ago

45 48 16 years ago

ngrams.googlelabs.com

goldins• 16 years ago

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.

_ouxp• 16 years ago

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.

gmartres• 16 years ago

Great! It just lacks an option for case-insensitive search.

programd• 16 years ago

There has been a ginormous increase in the use of the word "ginormous" in the last decade.

wgrover• 16 years ago

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?

tudorw2• 16 years ago

is good about to triumph over bad ?http://ngrams.googlelabs.com/graph?content=angel,devil&y...

tudorw2• 16 years ago

Man vs Machine http://ngrams.googlelabs.com/graph?content=PC,+secretary&...

rbancroft• 16 years ago

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.

igravious3• 16 years ago

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?

jluxenberg• 16 years ago

Also, "atomic": http://ngrams.googlelabs.com/graph?content=atomic&year_s...

akozak3• 16 years ago

This is really fun to play with: http://goo.gl/mNSDD

shib71• 16 years ago

Why would you use a shortener when there is no need? Especially when the URL is useful as it is here.

pkamb• 16 years ago

I think it was an atheist rick-roll.

akozak3• 16 years ago

It's fun to track how many clicks it gets, and I'd already had it on my clipboard for twitter :)

makmanalp• 16 years ago

http://ngrams.googlelabs.com/graph?content=christianity%2C+i...

akozak3• 16 years ago

It's actually case-sensitive, so here's a more accurate comparison: http://ngrams.googlelabs.com/graph?content=Christianity,Isla...

dchest• 16 years ago

"computer" and "internet" reveal Y2K problem.

staticshock• 16 years ago

i'm surprised that "star wars" has never been able to eclipse "galoshes":

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

Jun8• 16 years ago

Try "nerd". Like what you would expect. Now, try "sexy nerd". Sigh!

msbmsb5• 16 years ago

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".

igravious3• 16 years ago

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 :)

ivancho• 16 years ago

perfect example of http://en.wikipedia.org/wiki/Benford%27s_law

msbmsb5• 16 years ago

Exactly

msbmsb5• 16 years ago

Also interesting, the 1700s were not kind to the numbers "six" or "seven":

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

dwiel• 16 years ago

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 ...

msbmsb5• 16 years ago

The parsing of contractions/words with apostrophes is broken/unsupported:

http://ngrams.googlelabs.com/graph?content=can%27t%2Ccouldn%... vs http://ngrams.googlelabs.com/graph?content=t%2Ccan%27t%2Ccou...

igravious3• 16 years ago

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.)

ntoshev• 16 years ago

Google trends for books?

srean• 16 years ago

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.

nl6• 16 years ago

I love the "whale fail" image you get if you have an error: http://imgur.com/qCDCj

boredguy83• 16 years ago

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?

boredguy83• 16 years ago

wizard,magician,Harry Potter http://ngrams.googlelabs.com/graph?content=wizard,magician,H...

jpwagner3• 16 years ago

http://ngrams.googlelabs.com/graph?content=do+it+all+by+your...

jpwagner3• 16 years ago

http://ngrams.googlelabs.com/graph?content=can+you+help+me,l...

hanibash• 16 years ago

Someone please explain why "I love you" is cyclical???

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

nl6• 16 years ago

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.

nl6• 16 years ago

WTF? Why the downvote!?

jpwagner3• 16 years ago

wars and birthrates

ddlatham• 16 years ago

http://ngrams.googlelabs.com/graph?content=color,colour&...

Note the effects of the world wars.

jjclarkson• 16 years ago

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...

nl6• 16 years ago

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,...

ovi256• 16 years ago

What are those weird cycles in the Wall Street data ? Is that in any way related to the bubble/retraction cycle ?

nl6• 16 years ago

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.

zacharypinter• 16 years ago

Another fun one:

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

nl6• 16 years ago

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...

boredguy83• 16 years ago

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

msbmsb5• 16 years ago

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...

yoak• 16 years ago

Using American English throughout:

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

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

Google Labs: Books Ngram Viewer

Related Stories