This work was presented at the open repositories conference in Brisbane, Australia. The vast majority of digital data associated with scientific research are not accessible online. While there are many challenges associated with making research data openly accessible, one significant challenge is usability and long term availability of storage services. Open institutional repositories have the potential to support data preservation and sharing of valuable raw and processed data from local research efforts. However, research data are inherently heterogeneous and requires researcher involvement to accurately describe the nature of the deposited data files. We used a researcher-focused design principle to develop a data repository on the Invenio 3 platform with TIND. These principles included automating the deposit process as much as possible, employing standard metadata to support discoverability and future applications, and providing API access so the repository can power other visualization and analysis services. The repository includes DOI minting to support data citation, ORCID identifiers to facilitate credit attribution, and Github integration to encourage software archiving. The newly launched repository captures research data that might otherwise be lost due to poor storage and organization practices, and enables researchers, the library, and the Caltech Archives to develop tools and preservation strategies around this valuable resource.