Erhvervsudvalget 2022-23 (2. samling)
ERU Alm.del Bilag 164
Offentligt
2709247_0001.png
11 May 2023
2023 - 6017
cetono
Annex - Data access needs
a) What types of data, metadata, data governance documentation and other
information about data and how it is used can be useful to DSC’s for the
purpose of monitoring and assessing compliance and for vetted research-
ers for conducting research related to systemic risks and mitigation
measures?
Behavioral data:
-
Account creation date
-
Time spent logged in
-
Logins data/time
-
Clicks (articles, links, pictures, videos etc.) and with reference to
content including platform categorization (tags etc.)
-
Reactions (likes, shares, etc.)
–
tied to clicks and with reference to
content including platform categorization (tags etc.)
-
Time spent on individual items (articles, links, pictures, videos,
etc.)
-
Comments/posts (time/tags etc.)
-
Group associations
-
Followers/follows (or related concepts; friends etc.)
-
Purchases (if applicable)
-
Notifications (including user actions tied to notifications)
Demographic data (as defined by the platform). These can be explicit or
platform generated:
-
Age
-
Gender
-
Civil status
-
Sexual orientation
-
Political orientation
Algorithmic data:
-
Any commercial tags generated by the platform for individual users
(e.g. interests)
 ERU, Alm.del - 2022-23 (2. samling) - Bilag 164: Notat samt høringssvar vedr. adgang til meget store onlineplatforme og onlinesøgemaskiners data, fra erhvervsministeren
2709247_0002.png
2/2
-
-
-
-
Content recommendation tags/categories generated by the platform
for individual users
Content recommendation generated for the platform for specific
content (either content created on the platform, e.g. a post or tweet
as well as content generated outside the platform, e.g. an article,
video or similar).
Content sanction actions (deletion, warnings, clarifications etc.) for
specific content
Engagement scores generated by the platform for individual users
(if applicable)
Metadata:
-
Device data at user level
-
Referral flow data (data on entry
–
where is the user coming from,
and exit
–
where is the user going to, from the platform)
b) What sort of analysis and research might DSC's and vetted researchers
conduct for the purposes of monitoring and assessing compliance and con-
ducting research related to systemic risks and mitigation measures?
-
-
-
-
Descriptive analysis
–
measuring the magnitude of different varia-
bles
Correlational analysis
–
testing how different variables (e.g. age and
certain commercial tags) are associated
Time series analysis
–
analyzing how different variables (or rela-
tions between variables evolve over time)
Natural experiments
–
testing how different variables are causally
affected by others through regression discontinuity analysis or stag-
gered roll-out analysis.