EB-NeRD: Ekstra Bladet News
Recommendation Dataset

A Large-Scale Dataset for News Recommendation

Sponsored by


About

We present the Ekstra Bladet News Recommendation Dataset (EB-NeRD), a rich dataset released to support advancements and benchmarking in news recommendation research. It is collected from the user behavior logs of Ekstra Bladet, a classical legacy Danish newspaper published by JP/Politikens Media Group in Copenhagen. EB-NeRD comprises data from over 1 million unique users, with more than 37 million impression logs and over 251 million interactions from Ekstra Bladet. Alongside, we offer a collection of more than 125,000 news articles, enriched with textual content features such as titles, abstracts, and bodies. This enables text features in a low-resource language as context for recommender systems.

Dataset Description

RecSys 2024 Challenge

The dataset was released as part of the RecSys '24 Challenge, a competition that aims to advance news recommendation by addressing both the technical and normative challenges inherent in designing effective and responsible recommender systems for news publishing.
The RecSys '24 Challenge Workshop will be held at the 18th ACM Conference on Recommender Systems in Bari, Italy, from October 14-18, 2024. The workshop will take place on October 14, 2024. More details can be found here:

For more information about the dataset and the RecSys '24 Challenge, please refer to the following papers

EB-NeRD Paper
(Kruse et al. RecSys 2024)

Challenge Paper
(Kruse et al. RecSys 2024)


Download

EB-NeRD is free to download for research purposes under General License Terms (“License Terms”). Before you download the dataset, please read the License Terms and click below button to confirm that you agree to them.

I have read and accepted the License Terms

In addition, to help researchers become familiar with our data and run quick experiments, we are releasing a demo and a small version of the EB-NeRD by randomly sampling 5,000 and 50,000 users and their behavior logs from the full dataset.

Also, to get started, we have assembled a toolkit featuring a range of established news and general recommendation methods.

EBRec


Leaderboard

The evaluation system is hosted on the open-source platform Codabench. Please read the Challenge's Terms and Conditions. To register please follow:

Submit

The Codabench website is occasionally offline for maintenance to ensure optimal performance. To stay updated on maintenance schedules, please visit the CodaLab Competitions Google Group. Thank you for your understanding and continued support!


Updated: June 21, 2024.
To be added to the Academic Teams' leaderboard, please fill out the Google Form: Academic Leaderboard. Note, you should write your username for each member of the team, not the organization name in the form.

Search for multiple users at once using commas (e.g., "user_1, user_2, ..., user_n").
Rank User AUC MRR NDCG@5 NDCG@10