International Epilepsy Electrophysiology Portal

1. programmatically download or access data through filtering on annotations and metadata

Specifying annotations and values (say via XML) provides a direct download of custom data (with an XML legend to iterate over) or autogenerates a custom API for programmatically downloading slices

2. data uploader and data "structurer"

GUI and/or API for uploading data. Can specify fields which autogenerate a web view linked to specific files for entering metadata off of unstructured spreadsheets.

Upload a sample spreadsheet as template which is annotated once to create structure. Subsequent uploaded spreadsheets are automatically destructured and stored as metadata. Provide a spreadsheet template (column and row headers) to give to clinicians to facilitate uploading metadata.

3. "flip a switch" API. provide a base level API (e.g. custom parallel for-loops, data access, etc.) so that code can be run locally or on the cloud transparently with a flip of a switch, small test data locally, entire data set on the cloud.

4. data transformation tree

similar to source code version control, data transformations can be applied to uploaded data. both the original data and the new data are linked, with the version of the transformation code saved as well. people can navigate through different transformations of the data (filtering, FFT, feature extraction, etc.)

5. data comparison

web viewer can line up two datasets for comparison. can lock and unlock the scrolling so that the two datasets can be scrolled together or separately or carefully lined up and then scroll locked.

6. use statistics leader board (data downloads, most popular analyses), subscribe to data and code changes

7. code changes are versioned, source is visible, anyone can fork any code and upload changes and fixes (like github)