Pucktrick

Effortless data contamination for Data Centric AI studies

Learn more Feedback and Features request

An intuitive library for data contamination

Pucktrick is the ultimate library for managing introducing a controlled amount of specific errors into datasets trohugh providing intuitive methods for simulating real operating environment.

Friendly library for contaminating existing dataset

We know designing and implementing a data contamination systems can be a challenge and can demand both technical prowess and design skills, that is why we developed Pucktrick

Controlled contamination level

With Pucktrick, you can easily define the level of contamination that you want to inject into dataset

Wide range of real data error

with Pucktrick you will be able to simulate real error in your data, in order to recreate the real operating conditions of the datasets.