<s>
The	B-Application
Google	I-Application
Ngram	B-General_Concept
Viewer	I-General_Concept
or	O
Google	B-General_Concept
Books	I-General_Concept
Ngram	I-General_Concept
Viewer	I-General_Concept
is	O
an	O
online	B-Application
search	I-Application
engine	I-Application
that	O
charts	O
the	O
frequencies	O
of	O
any	O
set	O
of	O
search	O
strings	O
using	O
a	O
yearly	O
count	O
of	O
n-grams	B-Language
found	O
in	O
printed	O
sources	O
published	O
between	O
1500	O
and	O
2019	O
in	O
Google	B-Application
's	I-Application
text	O
corpora	O
in	O
English	O
,	O
Chinese	O
(	O
simplified	O
)	O
,	O
French	O
,	O
German	O
,	O
Hebrew	O
,	O
Italian	O
,	O
Russian	O
,	O
or	O
Spanish	O
.	O
</s>
<s>
The	O
n-grams	B-Language
are	O
matched	O
with	O
the	O
text	O
within	O
the	O
selected	O
corpus	O
,	O
optionally	O
using	O
case-sensitive	O
spelling	O
(	O
which	O
compares	O
the	O
exact	O
use	O
of	O
uppercase	O
letters	O
)	O
,	O
and	O
,	O
if	O
found	O
in	O
40	O
or	O
more	O
books	O
,	O
are	O
then	O
displayed	O
as	O
a	O
graph	O
.	O
</s>
<s>
The	B-Application
Google	I-Application
Ngram	B-General_Concept
Viewer	I-General_Concept
supports	O
searches	O
for	O
parts	O
of	O
speech	O
and	O
wildcards	B-Algorithm
.	O
</s>
<s>
It	O
was	O
inspired	O
by	O
a	O
prototype	O
called	O
Bookworm	O
created	O
by	O
Jean-Baptiste	O
Michel	O
and	O
Erez	O
Aiden	O
from	O
Harvard	O
's	O
Cultural	B-General_Concept
Observatory	I-General_Concept
,	O
Yuan	O
Shen	O
from	O
MIT	O
,	O
and	O
Steven	O
Pinker	O
.	O
</s>
<s>
The	O
Ngram	B-General_Concept
Viewer	I-General_Concept
was	O
initially	O
based	O
on	O
the	O
2009	O
edition	O
of	O
the	B-Application
Google	I-Application
Books	O
Ngram	B-Language
Corpus	O
.	O
</s>
<s>
The	O
Ngram	B-General_Concept
Viewer	I-General_Concept
returns	O
a	O
plotted	O
line	B-Application
chart	I-Application
.	O
</s>
<s>
Due	O
to	O
limitations	O
on	O
the	O
size	O
of	O
the	O
Ngram	B-Language
database	O
,	O
only	O
matches	O
found	O
in	O
at	O
least	O
40	O
books	O
are	O
indexed	O
in	O
the	O
database	O
.	O
</s>
<s>
The	O
data	O
set	O
has	O
been	O
criticized	O
for	O
its	O
reliance	O
upon	O
inaccurate	O
OCR	B-Application
,	O
an	O
overabundance	O
of	O
scientific	O
literature	O
,	O
and	O
for	O
including	O
large	O
numbers	O
of	O
incorrectly	O
dated	O
and	O
categorized	O
texts	O
.	O
</s>
<s>
Guidelines	O
for	O
doing	O
research	O
with	O
data	O
from	O
Google	B-General_Concept
Ngram	I-General_Concept
have	O
been	O
proposed	O
that	O
address	O
many	O
of	O
the	O
issues	O
discussed	O
above	O
.	O
</s>
<s>
Optical	B-Application
character	I-Application
recognition	I-Application
,	O
or	O
OCR	B-Application
,	O
is	O
not	O
always	O
reliable	O
,	O
and	O
some	O
characters	O
may	O
not	O
be	O
scanned	O
correctly	O
.	O
</s>
<s>
Although	O
Google	B-General_Concept
Ngram	I-General_Concept
Viewer	I-General_Concept
claims	O
that	O
the	O
results	O
are	O
reliable	O
from	O
1800	O
onwards	O
,	O
poor	O
OCR	B-Application
and	O
insufficient	O
data	O
mean	O
that	O
frequencies	O
given	O
for	O
languages	O
such	O
as	O
Chinese	O
may	O
only	O
be	O
accurate	O
from	O
1970	O
onward	O
,	O
with	O
earlier	O
parts	O
of	O
the	O
corpus	O
showing	O
no	O
results	O
at	O
all	O
for	O
common	O
terms	O
,	O
and	O
data	O
for	O
some	O
years	O
containing	O
more	O
than	O
50%	O
noise	O
.	O
</s>
