The goal of this project is to improve accessibility of open datasets by curating them. “NiData” aims to provide a common interface for documentation, downloads, and examples to all open neuroimaging datasets, making data usable for experts and non-experts alike.
NiData is a Python package that provides a single interface accessing data from a variety of open data sources. The software framework makes it easy to add new data sources, simple to define and to provide access to multiple datasets from a single data source. Software dependencies are managed on a per-dataset basis, allowing downloads and examples to use any public packages without requiring installation of packages required by unused datasets. The interface also allows selective download of data (by subject or type) and caches files locally, allowing easy management of big datasets.
We focused on exposing new methods for downloading data from the HCP, supporting access via Amazon S3 and HTTP/XNAT. We were able to provide a downloader that accepts login credentials and downloads files locally. We created an example that interacts with DIPY to produce diffusion imaging results on a single subject from the HCP. We also worked at collecting common data sources, as well as individual datasets stored at each data source, into NiData’s “data sources” wiki page. We incorporated downloads, documentation, and examples from the nilearn package and began discussion of making a more extensible object model.
Read the peer-reviewed publication(s):