Session description
Since CoSector’s recent eprints repository implementation our aim was to bring our repository into line with our strategic goal to foster LSE open research and upgrade our platform for the 21st Century. Ensuring our technical infrastructure was robust and secure was a key driver, as well as the ability to showcase all of LSE’s research outputs in a modern and user-friendly platform. This case study will explore the project from initiation through to completion and highlight how a close working relationship with Cosector was key to its success.
Session recording
Loading PDF…
Unable to load this PDF.
Optional PDF Caption.
Session summary
LSE had long relied on its repository—“over 100,000 records, 60,000 full text, 30 million downloads worldwide”—as a central tool for public engagement. However, the legacy ePrints system had become outdated, difficult to maintain, and poorly aligned with modern expectations. As one speaker noted, the old interface “looked like 2004… didn’t display very well on mobiles… [and] wasn’t really meeting our researcher needs.” The shift of IT support to central services meant the library no longer had in‑house expertise to maintain ePrints, making a sustainable externally supported solution necessary.
A new institutional research strategy and the broader Research Outputs and Discovery (ROAD) programme created the opportunity to rethink workflows for publications, theses, and research data. LSE had never operated an institutional data repository, historically directing researchers to the UK Data Archive. But the school’s renewed commitment to open social science required a more comprehensive offering. The team emphasised the need for tailored support for sensitive social science data and the creation of a data catalogue to track externally deposited datasets.
Co‑Sector was selected after an extensive procurement process because it best met LSE’s functional requirements, offered flexibility, and had strong experience integrating ePrints with Pure. The project required significant technical development, including a new dataset connector using the Pure API. Because Pure could not natively handle dataset records in the same way as publications, the team developed a bespoke workflow in which metadata was managed in Pure while files were stored in ePrints. This required innovations such as generating “dummy files” in Pure to preserve metadata consistency.
Other major technical work included redesigning metadata structures, implementing Pure UUIDs for reliable person identification, restructuring the subject taxonomy, and improving performance by generating browse views and statistics overnight rather than on demand.
The migration of LSE’s separate theses repository into the new unified system required careful handling of overlapping record IDs. The team exported the old database, transformed records into Pure‑compatible XML, and synchronised them into the new ePrints instance while preserving legacy identifiers and redirects.
A major focus of the project was the new public interface. Extensive iterative testing with staff and researchers led to redesigned search, improved person browse pages, and new “splash pages” for theses and datasets to ensure these smaller collections remained visible. A new statistics dashboard visualised global reach, including a world map showing download activity—described as a “simple but impactful visualisation” that highlighted the repository’s global reach.
The project was completed in an intense four‑to‑five‑month window, supported by strong communication between LSE and Co‑Sector and an “invaluable” project manager. While the launch delivered a complete, functional service, further work was planned, including dataset versioning, larger file support, mobile optimisation, search improvements, and integration with Preservica for preservation.