Skip to content

Dataset · criteo

Large Criteo dataset for display advertising attribution

criteo/criteo-attribution-dataset

From 10 to 100 million tabular records from Criteo Research for attribution models and ad bidding.

What's inside

The tabular dataset for a binary classification task includes 10 to 100 million rows. The dataset was published by Criteo Research and Facebook researchers alongside a KDD 2017 research paper on improving display advertising bidding efficiency using attribution modeling. Detailed feature composition and storage format are not described at the beginning of the card.

What you can do with it

  • Training tabular models for attribution and conversion evaluation in ad networks
  • Benchmarking gradient boosting and neural network algorithms on large tabular data
Tasks
Tabular classification
Size
10M–100M rows
Downloads this month
25.6K
Likes
12

How to load it

from datasets import load_dataset

ds = load_dataset("criteo/criteo-attribution-dataset")