University of Maryland researcher maps Southern Maryland’s innovation networks

Dr. Scott Dempelt of the University of Maryland presented an innovation-network analysis to the Saint Mary's County Economic Development Commission on Dec. 17, showing how patent and research-grant data can reveal real-time links among inventors, companies and technologies in Southern Maryland.

Why it matters: Dempelt’s method maps people and organizations as nodes and their patent- and grant-based relationships as links. The approach aims to identify a network’s core and brokerage nodes so county staff can target specific individuals, firms and technologies for commercialization, licensing or local business development.

What Dempelt showed

Dempelt said conventional cluster analysis lags innovation because employment and industry code data are generated years after invention events. “If you’re in an innovation driven economy and you’re not getting information on the innovation drivers until 5 years later, that’s not gonna work,” he told the commission.

He described building weekly-updated networks from patent applications, patent grants and research-grant data, creating small networks from each document and layering them to form a large, complex graph. Nodes represent inventors, organizations and locations; ties represent co-inventorship, assignees and shared technology classes. Dempelt emphasized the distinction between a network’s highly connected core—where serial inventors and frequent collaborators cluster—and the periphery, where single, possibly unrelated filings appear.

Dempelt demonstrated a regional network with a prominent Navy-centered core and clusters tied to Montgomery County and to the College of Southern Maryland. He highlighted several local inventors and brokerage figures who link Pax River-related activity to outside firms and research organizations; attendees noted licensing and commercialization opportunities often go to firms outside the county.

Interactive tool and next steps

Dempelt said the analysis is being automated and an interactive model (built as an open-source Excel plug-in) will be delivered to county staff. Planned deliverables include a draft analytic report, regular network updates and a user training session in January or February. He told the commission that licensing, sponsored-research and private-investment data can be added to the model under confidentiality protocols.

Commission members asked whether the tool could help identify non‑defense commercialization (for example, aquaculture), and Dempelt said the platform can incorporate additional data sources (research grants, Crunchbase/AngelList investment data, USDA information) to surface cross-sector opportunities. He also noted that while each patent tie receives an equal weight in the network model, users can add licensing or investment data later to reflect economic value.

Direct quote and analogy

When asked whether the tool could help ‘matchmake’ firms and inventors, a meeting participant said the model looked like “eHarmony for inventors,” an observation Dempelt endorsed as a useful way to think about targeted outreach to specific people and technologies.

Ending

Dempelt told the commission he will return with updated data and a draft report; staff will receive the interactive model and training. Commissioners and staff said they plan to use the network maps to prioritize outreach, refine the SEDS work and explore whether local licensing and commercialization can be captured in-county rather than by outside firms.