license: cc-by-nc-4.0
pretty_name: MerRec
size_categories:
- 1B<n<10B
language:
- en
tags:
- recommendation
- sequential recommendation
- click-through rate prediction
- e-commerce
MerRec: A Large-scale Multipurpose Mercari Dataset for Consumer-to-Consumer Recommendation Systems
This repository contains the dataset accompanying the paper MerRec: A Large-scale Multipurpose Mercari Dataset for Consumer-to-Consumer Recommendation Systems at KDD 2025.
Contributors: Lichi Li, Zainul Abi Din, Zhen Tan, Sam London, Tianlong Chen, Ajay Daptardar
Overview
The MerRec dataset is a large-scale, highly diverse, thoroughly anonymized and derived subset of item interaction event sequence data from Mercari, the C2C marketplace e-commerce platform. It is designed for researchers to study recommendation related tasks on a rich C2C environment with many item features.
Some basic statistics are:
- Unique users: Over 5 million
- Unique items: Over 80 million
- Unique events: Over 1 billion
- Unique sessions: Over 200 million
- Item title text tokens: Over 8 billion
For a detailed walkthrough and an extensive list of accurate statistics, feature interpretations, preprocessing procedure, please refer to the paper.
File Organization
The MerRec dataset is divided into 6 directories, each containing about 300 Parquet shards from a particular month in 2023.
Experiments
Code implementation used for the experiment section of the paper can be found here.
BibTeX
If you found our work useful, please consider citing MerRec:
@inproceedings{10.1145/3690624.3709394,
author = {Li, Lichi and Din, Zainul Abi and Tan, Zhen and London, Sam and Chen, Tianlong and Daptardar, Ajay},
title = {MerRec: A Large-scale Multipurpose Mercari Dataset for Consumer-to-Consumer Recommendation Systems},
year = {2025},
isbn = {9798400712456},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3690624.3709394},
doi = {10.1145/3690624.3709394},
abstract = {In the evolving e-commerce field, recommendation systems crucially shape user experience and engagement. The rise of Consumer-to-Consumer (C2C) recommendation systems, noted for their flexibility and ease of access for customer vendors, marks a significant trend. However, the academic focus remains largely on Business-to-Consumer (B2C) models, leaving a gap filled by the limited C2C recommendation datasets that lack in item attributes, user diversity, and scale. The intricacy of C2C recommendation systems is further accentuated by the dual roles users assume as both sellers and buyers, introducing a spectrum of less uniform and varied inputs. Addressing this, we introduce MerRec, the first large-scale dataset specifically for C2C recommendations, sourced from the Mercari e-commerce platform, covering millions of users and products over 6 months in 2023. MerRec not only includes standard features such as user_id, item_id, and session_id, but also unique elements like timestamped action types, product taxonomy, and textual product attributes, offering a comprehensive dataset for research. This dataset, extensively evaluated across three recommendation tasks, establishes a new benchmark for the development of advanced recommendation algorithms in real-world scenarios, bridging the gap between academia and industry and propelling the study of C2C recommendations. Experiment code (https://github.com/mercari/mercari-ml-merrec-pub-us) and dataset (https://huggingface.co/datasets/mercari-us/merrec) are released.},
booktitle = {Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1},
pages = {2371–2382},
numpages = {12},
keywords = {datasets, recommender systems},
location = {Toronto ON, Canada},
series = {KDD '25}
}
License
Dataset license: CC BY-NC 4.0 International