<s>
Culturomics	B-General_Concept
is	O
a	O
form	O
of	O
computational	O
lexicology	O
that	O
studies	O
human	O
behavior	O
and	O
cultural	O
trends	O
through	O
the	O
quantitative	O
analysis	O
of	O
digitized	O
texts	O
.	O
</s>
<s>
Researchers	O
data	B-Application
mine	I-Application
large	O
digital	O
archives	O
to	O
investigate	O
cultural	O
phenomena	O
reflected	O
in	O
language	O
and	O
word	O
usage	O
.	O
</s>
<s>
Michel	O
and	O
Aiden	O
helped	O
create	O
the	O
Google	B-Application
Labs	I-Application
project	O
Google	B-General_Concept
Ngram	I-General_Concept
Viewer	I-General_Concept
which	O
uses	O
n-grams	B-Language
to	O
analyze	O
the	O
Google	B-Application
Books	I-Application
digital	O
library	O
for	O
cultural	O
patterns	O
in	O
language	O
use	O
over	O
time	O
.	O
</s>
<s>
Because	O
the	O
Google	B-General_Concept
Ngram	I-General_Concept
data	O
set	O
is	O
not	O
an	O
unbiased	O
sample	O
,	O
and	O
does	O
not	O
include	O
metadata	O
,	O
there	O
are	O
several	O
pitfalls	O
when	O
using	O
it	O
to	O
study	O
language	O
or	O
the	O
popularity	O
of	O
terms	O
.	O
</s>
<s>
In	O
a	O
study	O
called	O
Culturomics	B-General_Concept
2.0	O
,	O
Kalev	O
H	O
.	O
Leetaru	O
examined	O
news	O
archives	O
including	O
print	O
and	O
broadcast	O
media	O
(	O
television	O
and	O
radio	O
transcripts	O
)	O
for	O
words	O
that	O
imparted	O
tone	O
or	O
"	O
mood	O
"	O
as	O
well	O
as	O
geographic	O
data	O
.	O
</s>
<s>
The	O
authors	O
also	O
identified	O
a	O
universal	O
"	O
tipping	O
point	O
"	O
in	O
the	O
life	O
cycle	O
of	O
new	O
words	O
at	O
about	O
30	O
to	O
50	O
years	O
after	O
their	O
origin	O
,	O
they	O
either	O
enter	O
the	O
long-term	O
lexicon	B-Application
or	O
fall	O
into	O
disuse	O
.	O
</s>
<s>
This	O
study	O
was	O
performed	O
on	O
a	O
vast	O
scale	O
,	O
by	O
using	O
statistical	B-General_Concept
machine	I-General_Concept
translation	I-General_Concept
,	O
text	B-Algorithm
categorisation	I-Algorithm
and	O
information	B-General_Concept
extraction	I-General_Concept
techniques	O
.	O
</s>
<s>
The	O
possibility	O
to	O
detect	O
mood	O
shifts	O
in	O
a	O
vast	O
population	O
by	O
analysing	O
Twitter	B-Application
content	O
was	O
demonstrated	O
in	O
a	O
study	O
by	O
T	O
.	O
Lansdall-Welfare	O
and	O
co-authors	O
.	O
</s>
<s>
In	O
2015	O
,	O
a	O
study	O
revealed	O
the	O
bias	O
of	O
the	O
Google	B-Application
books	I-Application
data	O
set	O
,	O
which	O
"	O
suffers	O
from	O
a	O
number	O
of	O
limitations	O
which	O
make	O
it	O
an	O
obscure	O
mask	O
of	O
cultural	O
popularity	O
,	O
"	O
and	O
calls	O
into	O
question	O
the	O
significance	O
of	O
many	O
of	O
the	O
earlier	O
results	O
.	O
</s>
<s>
In	O
2016	O
,	O
a	O
publication	O
by	O
Richard	O
Ladle	O
and	O
colleagues	O
highlighted	O
five	O
key	O
areas	O
where	O
culturomics	B-General_Concept
can	O
be	O
used	O
to	O
advance	O
the	O
practice	O
and	O
science	O
of	O
conservation	O
,	O
including	O
recognizing	O
conservation-oriented	O
constituencies	O
and	O
demonstrating	O
public	O
interest	O
in	O
nature	O
,	O
identifying	O
conservation	O
emblems	O
,	O
providing	O
new	O
metrics	O
and	O
tools	O
for	O
near-real-time	O
environmental	O
monitoring	O
and	O
to	O
support	O
conservation	O
decision	O
making	O
,	O
assessing	O
the	O
cultural	O
impact	O
of	O
conservation	O
interventions	O
,	O
and	O
framing	O
conservation	O
issues	O
and	O
promoting	O
public	O
understanding	O
.	O
</s>
<s>
Others	O
have	O
demonstrated	O
bias	O
in	O
the	O
Ngram	B-Language
data	O
set	O
.	O
</s>
<s>
Their	O
results	O
"	O
call	O
into	O
question	O
the	O
vast	O
majority	O
of	O
existing	O
claims	O
drawn	O
from	O
the	O
Google	B-Application
Books	I-Application
corpus	O
"	O
:	O
"	O
Instead	O
of	O
speaking	O
about	O
general	O
linguistic	O
or	O
cultural	O
change	O
,	O
it	O
seems	O
to	O
be	O
preferable	O
to	O
explicitly	O
restrict	O
the	O
results	O
to	O
linguistic	O
or	O
cultural	O
change	O
‘	O
as	O
it	O
is	O
represented	O
in	O
the	O
Google	B-General_Concept
Ngram	I-General_Concept
data’	O
"	O
because	O
it	O
is	O
unclear	O
what	O
caused	O
the	O
observed	O
change	O
in	O
the	O
sample	O
.	O
</s>
