api or archives?
tormented ways to transform tweets
into historical sources

frédéric clavert / @inactinique /

15 Septembre 2022

historian / #ww1 / #covid

illustration: projet esch-mars / simon beuret

archives?

archives? (in french)

archiving as the setting of hisorians’ work

preserving / sorting / indexing

primary sources and archives


all archives are primary sources

all primary sources are not necessarily archives

tweets as primary sources

  • twitter’s api as a way to get primary sources, not archives
  • twitter’s api as settings of a researcher’s work

api as settings of the historians’ work

is there an api? / what are the conditions to use the api? / sustainability of the api

researching echoes of the past on twitter

#ww1 (api 1.1)

#covid (api 1.1, apiv2)

#ww1 and the echoes of the centenary of the first world war online

  • 2014-2019
  • 9 million tweets
  • 1.5 Twitter account

api as a framework

  • twitter streaming api (v1.1)
  • anticipation
  • no going back

some results

covid19 on twitter

  • with Deborah Paci
  • massive harvesting of tweets, first in french, then in italian too (comparison)
  • started with a harvesting of the same nature as the #ww1 project
  • 2021: API v2

new api, new framework…

  • 10 million tweets
  • history of twitter
  • are we closer to an archive?

…and other limitations

working on archived tweets

what would be the point?

  • I work with CSVs, JSONs not with what you actually see, read on twitter
  • the environment of a tweet is not historicized if you use the API

an example: Obama’s ‘4 more years’ tweet

and in 2012 (wayback machine)

as a conclusion: how to really historicize twitter?

images credits