Organize dataset

Background

All SPARC datasets must follow the top level SPARC folder structure imposed by the SPARC Dataset Structure. This top level folder structure is shown in the figure below. If your data organization doesn't follow this structure inherently, you can create it virtually with SODA then either generate it locally on your computer or directly on Blackfynn (to avoid duplicating files locally).

Overview of the top level folder structure required for all SPARC datasets (taken from the instructions prepared by the SPARC Curation Team).

How to

Step 1: Getting started

Step 2: Specify high-level folders

Step 3: Structure dataset files

Step 4: Specify high-level metadata files

Step 5: Request manifest files

Step 6: Generate dataset

Step 7: Preview dataset

Note that you can save your progress using the Save progress button available in the lower right corner starting from Step 3.