Customer's have used the document below to highlight Flywheel's capabilities in grant applications. Edit this document as needed to fit your institution's use cases and application.
The Flywheel scientific data and algorithm management system satisfies IRB and funding sources’ data collection, protection, and collaboration guidelines and requirements.
The Flywheel data and algorithm management system is a centralized database that enables collaboration and data access to image-intensive datasets, as well as management of the algorithms used to generate the data. Additionally, the solution enables researchers to accelerate their research by removing the information technology (IT) burden of content management.
The Flywheel solution will be used to share data in compliance with various funding sources’ data sharing guidelines(e.g. NIH). As the day-to-day data management tool for this project, Flywheel will allow researchers to set permissions on the data so that it can be accessed privately or publicly both efficiently and quickly.
Flywheel’s system is an API architecture that handles all data going in and out of the platform through a single, secure, permissions-based model. Data is encrypted while in transit using SSL certificates and can also be encrypted at rest. All data is managed in a secure project context with access controls integrated with institutional authentication protocols.
Data collection will be handled securely and efficiently using Flywheel’s modality integration software, folder organizer and uploading tool for existing data, and web application uploading tool. More specifically, data is ingested at the original point of data acquisition.
The data that is collected will be indexed into a database and organized by pre-determined projects. The database is integrated with Elasticsearch which allow users to search for data types, acquisition types, analysis outputs, annotated fields, and tags generated within the platform.
Researchers can access data via Flywheel’s web-interface or leverage its command-line-interface (CLI) and software-development-kits (SDKs). All the data access modules can be installed by initially generating an API Key from the Flywheel web-interface. Users are provided with extensive documentation for all the data access options.
Flywheel also permits users to create their own analysis methods, referred to as “gears” in the product, while following proper reproducible research practices. Additionally, gears on the platform can be configured per project to automatically trigger or link together with each other to create pipelines on the web-interface or via CLI or SDKs. The gears can be run on on-premises or in a scalable fashion in the cloud. Each gear, when deployed, can be configured to run on multiple virtual machines (VMs) in the cloud.
Datasets that are ready to be made public will move to a Flywheel collection that is available for download. This collection will act as a long-term data archive. All data will be de-identified before being made public and will include proper documentation by way of attachment.
The analysis software tools found in Flywheel provide reproducible calculations on all data within the system. Using pre-packaged analyses (referred to as “gears”), the platform will enable high-quality, reproducible research and accurate recording of all completed analysis. The recording of all run gears (success and failure indications), results, and workflows are documented on the web-interface or easily accessed via CLI.