9M44hALYt S. LAO - NMT U.MBER
ANALYSIS & NUMBERS
NN
I
N
STARTS ARCHIVING
PUBLIC DANISH FACEBOOK
I
cg-a
C.
Thu
Facebook is the dominant social media in Denmark and is regarded as
Denmark.
the main arena for online public discourse in
AWM
/
Analysis & Numbers starts collecting all posts and comments
Facebook.
from all Danish political and media pages on
Ccp-c
14
I<
* U
We've gained access to the Facebook API by creating a
couple of Facebook"apps.
LL-L+ A
D EVE LCi
)
* They get some system-level tokens that are allowed d
to viewpublic information on larger Facebook pages
(>1000 followers), but with all user information (user names and IDs)
scrubbed.
2~
91. R 91'
/
* The colle ction program is written in Python and data is saved to
local Kubernetes cluster.
a replicated Postgres database running in our
wwrwanidnumbersxcom.
* For the Facebook groups, we use a large collection
facebook.com.
of Facebook profiles that scrape the mobile version of
* These profiles are run in containers that use proxies and HTTP user
different 4G masts.
agents that place them as coming from
* This creates good IP rotation while not seeming unnatural,
blocking.
and therefore less suspicious and susceptible to
* However, this is still rather fragile, as Facebook does not
rate limit is low but unknown.
shy away from blocking or deactivating users, and the
The rise of 50...
Extraction of .
Tt MutLtitorc Netarkivet - t
sq
i
We bdata - Bu
50Lroayback - Advarced search-, & discovery
Analysis & ...
Heritrix pro...
2014
[ First mappi...
50 The Nation... Denmarkc 50 The Archivo.
i me se
rbridqe An..
20.9
?0 0
200ii5
9
'N
I
1008
1
2010
201 1
1
1015
*6
*5 asa
2017
em
2020
I2 U
1022
* ISe
20 23
200
2 r..
202-i
:') 25
2016
