Drug Listing Dataset

Gwern published a collection of Darknet Market Archives in 2015. While the dataset is comprehensive, a lot of works need to be done before we can start analyzing it. I put some effort to organize Gwern’s archives, extract only drug-related content, and construct a drug listing dataset in csv format.

Feel free to check out my repo here.

Coverage

Drug listing data are parsed from the following marketplaces.

  • 1776
  • Abraxas
  • Agora
  • Evolution
  • Nucleus
  • Outlaw Market
  • Silk Road 2
  • The Marketplace

Data Field

  • product_title: The title of the item.
  • product_description: The description of the item. This field could be null as some of the listing archives are damaged.
  • ship_from: The place where the item is shipped from. This field could be null as shipping information isn’t required for certain marketplaces.
  • ship_to: The place where the item is shipped to. This field could be null as shipping information isn’t required for certain marketplaces.
  • seller: The seller of the item.
  • price: The price of the item, the currency is not yet unified.
  • source: The name of the marketplace where this item is posted.

Possible Research

Drug Slang Detection

Wanzheng et al. (2021) proposed an unsupervised method to detect and identify drug slang. By masking a drug name (e.g. heroin), they use a language model to fill in the mask with words other than the original drug name. Those predicted words are considered as drug slang. This dataset could be used to enhance their language model.