Managing and organizing large data sets can be a daunting challenge. In our lab, which focuses on the development of devices to restore function after damage to the nervous system, experiments may include tens or hundreds of trials, and individual trials may generate multiple large data files. Further, each experiment may involve a unique hardware setup that can impact the structure, meaning, and quality of the data collected. To allow us to characterize, manage, organize, and search these large data sets, we have developed a Drupal-based management system to capture and store all metadata about each experiment.

This metadata database (MDDB) allows for a reliable and standard way to collect metadata to improve experimental replication, reduce error sources due to missing data, and provide new ways to organize and analyze data sets across trials and experiments.

The MDDB project has a wide array of advantages for managing large data sets, including:
- platform independence
- minimal deployment requirements
- ease of extendibility
- flexibility
- ability to export data in many formats
- ability to collect data from a GUI or programmatically
- clear separation between data structure and presentation, and
- a reduced number of clicks when entering content, improving functionality during experimental sessions.

Using content types, theming, web services, views, contributed and costume modules, we developed a functional MDDB system in a relatively short period of time. While the system is always under constant development, the production version has been exceptionally stable during many recent experiments.

This session will provide insight into the project rationale, the process from development through production using continuous integration, and future directions that will include automated manipulation of newly generated data sets so that they can be imported into a separate analysis framework.