Dataset for the "Give and Take" Paper

This page accompanies the paper "Give and Take: An End-To-End Investigation of Giveaway Scam Conversion Rates", published at ACM IMC 2024. Below, you will find details about the dataset, including a link to download it and citation information.

Download

You can download the dataset as a ZIP file from the link below:

Download Dataset (ZIP, 7.9GB)

This ZIP file is password-protected. Please email e7liu@ucsd.edu to request the password.

What's in the Dataset

The dataset only contains YouTube-related data, including 2-second samples of YouTube livestreams (and their associated chat messages), scam websites advertised by the livestreams, screenshots and HTML of the scam websites, and addresses advertised on those websites. Specifically, the above zip file contains three folders and a summary file:

• videos: Videos organized by video id (vid). Each subdirectory is named with the following format: vid_<the_actual_vid>.
• chats: Chat messages associated with each video, also organized by video id.
• websites: Screenshots and HTML of each website organized by the scam website's URL.
• websites_and_data.txt: Summary of information organized by each scam website, including details such as livestreams advertising the website and cryptocurrency addresses listed on it.

Citation

If you use this dataset in your research, please cite our paper:

@inproceedings{liu2024give,
  title={Give and take: An end-to-end investigation of giveaway scam conversion rates},
  author={Liu, Enze and Kappos, George and Mugnier, Eric and Invernizzi, Luca and Savage, Stefan and Tao, David and Thomas, Kurt and Voelker, Geoffrey M and Meiklejohn, Sarah},
  booktitle={Proceedings of the 2024 ACM on Internet Measurement Conference},
  pages={704--712},
  year={2024}
}