Data must be open--public, easy to access, and well-documented--to be scientific. As a data analyst interested in examining brain images, I found it virtually impossible to locate datasets, let alone download them, understand what data was contained, and figure out how to processess them.

I created the open-source nidata project to accomplish these goals. The project aims to aggregate all the different databases where data exist, provide each with a simple method to access the data (downloads if necessary), and provides documentation with a working example on how to use the data.

The code works from code developed in the nilearn project, which uses machine learning to analyze neuroimaging data. I have extended the code, added new databases, and contributed back code to nilearn whenever possible.