Tracker Insights
last updated: 15. November 2024
Decision Overview
Thus far, 31 complete applications have been registered through our tracker.
We know that 24 have been decided by the platforms.
Out of the decided applications, 15 have been accepted and 9 rejected.
We can introduce more nuance by differentiating our data by platforms: applications to Facebook, TikTok, YouTube, X.com, and Bing have been shared with us, with X.com and TikTok making up the majority of our dataset.
| Platform | Data Requested | Status | # |
|---|---|---|---|
| X.com | public data | open | 4 |
| rejected | 8 | ||
| accepted | 5 | ||
| TikTok | public data | open | 1 |
| rejected | 1 | ||
| accepted | 9 | ||
| YouTube | public data | accepted | 1 |
| public data | open | 1 | |
| Bing | non-public data | open | 1 |
Below you can see the distribution of time between application and decision for X.com and TikTok. The mean time differences are shown below.
TikTok
all decisions
37.5 days
time to reject
81 days
time to accept
32.67 days
X.com
all decisions
71.23 days
time to reject
43.13 days
time to accept
116.2 days
Reasons for Rejection
"Based on your application, it does not appear that your proposed use of data is solely for performing research that contributes to the detection, identification and understanding of systemic risks in the EU as described by Art. 34 of the Digital Services Act."
The above statement is the most common reason for rejection we find in for the applications in our dataset (n=4).
Other reasons include the incorrect conditions that you must be affiliated with a university (n=1) or located inside the European Union (n=2).
Two other application were denied because of incompleteness (n=1) and because the access request deviated from the application (n=1).
Application Insights
The applications we received thus far cover a range of topics.
Methods
The most popular methods, mentioned by 5 applications, are (not closer specified) machine learning approaches and network analysis. Four applications also reference Natural Language Processing (NLP, including topic modelling or sentiment analysis). Content analysis is mentioned by three applications. Three more applications also plan to combine survey and interview data with the data received through data access. Two applications mention deep learning-based approaches for multi-modal analysis.
Risks
Since data access is only granted for „research that contributes to the detection, identification and understanding of systemic risks in the Union“ many applications mention the risks specified in Art. 34(1) DSA,:
- Dissemination of Illegal Content (n=4)
- Negative Effects on Fundamental Rights (n=3)
- Negative Effects on Civic Discourse, Electoral Process & Public Security (n=3)
- Negative Effects on Gender-Based Violence, Public Health, Minors & Well-being (=2)
However, only 5 out of 16 entries that provided their application text made specific reference to Art. 34.