← Projects

Florida Ocean Economy Database

How the dataset is defined, collected, classified and de-duplicated

Purpose & Approach

This methodology documents the steps taken to build a dataset of emerging ocean economy companies in Florida. This project is a personal exercise in learning about the ocean economy landscape in Florida. Using publicly available sources and with the assistance of Claude, I have scraped online sources for companies, categorized them by sector and innovation theme, and compiled summary information on each one. The goal will be to continue bringing in new data sources as I find them, and to explore ways to make this database useful to the Florida community.

The methodology is laid out in the following sections:

Data Sources

No single source captures the ocean economy. Companies surface through different channels at different stages — federal grant databases catch R&D-stage firms, while local accelerators and award programs catch ventures earlier, often before they have any funding record at all. The dataset therefore combines multiple discovery channels, with each new source parsed into the same standardized company-level schema so that everything can be merged, deduplicated, and compared.

Federal Awards: SBIR/STTR

The U.S. federal government offers Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) awards to emerging startups. Data on SBIR/STTR awards is published on a quarterly basis, and includes information such as award abstracts, research keywords, and activity descriptions that make the dataset convenient to analyze.

Search process: the full national dataset is downloaded from SBIR.gov via its API. The following filters were applied:

Federal AgencyIncluded Branches
Department of Agriculture (USDA)Incl. National Institute of Food and Agriculture (NIFA)
Department of Commerce (DOC)National Oceanic and Atmospheric Administration (NOAA); National Institute of Standards and Technology (NIST)
Department of Defense (DOD)Incl. Navy and Defense Advanced Research Projects Agency (DARPA)
Department of Energy (DOE)Incl. ARPA-E
Department of Homeland Security (DHS)Science and Technology Directorate
Department of Transportation (DOT)All branches
Environmental Protection Agency (EPA)All branches
National Science Foundation (NSF)All branches

Florida-Based Organizations

Several start-up accelerators and incubators operate in Florida, including Seaworthy Collective and Ocean Exchange. Each organization publishes information about the startups that it has supported. Web searches were conducted to scrape information from these organizations.

Additional programs are added as they are identified — including The Continuum, a NOAA-funded national network of ocean-economy accelerators (Tampa Bay Wave's BlueTech|X and Braid Theory cohorts, among others), and FAU Tech Runway's pitch-competition awards. Cohort lists, award announcements, and alumni pages are parsed into the standard schema, with a source URL retained for every record.

Classification

Companies are classified using Anthropic's Claude AI. Classification is driven by the source text available for each company — award abstracts and research keywords for federally funded companies, program descriptions and cohort announcements for accelerator-sourced companies. Where the source text is ambiguous, the company website is reviewed for evidence of the company's primary line of business.

Why AI classification?

Each company is classified along two dimensions: ocean economy sector and innovation theme. The Sector reflects the established industry in which the company operates; the Innovation Theme reflects the emerging blue-economy activity it represents. A company can sit in an established sector and carry an emerging theme — an eco-engineered seawall venture, for example, is Construction sector + Coastal Resilience theme.

Ocean Economy Sectors

Sectors are aligned to FAU's Florida ocean-economy industry framework, keeping the dataset comparable to FAU's published economic baseline for the state.

CategoryDescription
ConstructionCoastal protection and resilient marine infrastructure — eco-engineered shorelines, 3D-printed coastal structures, flood barriers
EnergyMarine renewable energy — wave energy converters and grid-scale marine energy systems
Living ResourcesAquaculture, fisheries, and marine biotechnology — broodstock systems, species identification, sea-lice removal
Marine IndustryVessel operations and maintenance — hull grooming, cleaning robotics, and shipboard systems
MilitaryNaval and undersea defense technology — sonar, undersea sensing, and ship systems
Ocean IntelligenceOcean observation and data — autonomous underwater vehicles, in-situ sensor networks, underwater LiDAR

Innovation Themes

Established industry frameworks report jobs and GDP but have no home for emerging blue tech such as blue carbon, marine-plastics remediation, or ecosystem restoration. The Innovation Theme layer surfaces these emerging niches — exactly the untracked segments this project targets. Ventures with no established industry home (e.g., blue-carbon fintech) are tagged Cross-Sector / Enabling Tech at the sector level, and their Innovation Theme carries the real signal.

ThemeDescription
Ocean Data & IntelligenceOcean sensing, observation, autonomous platforms, and marine data & analytics
Sustainable MaterialsBio-based, recycled, and ocean-safe alternatives to conventional materials
Ecosystem Restoration & ConservationCoral, reef, and coastal habitat restoration and conservation technology
Regenerative Ocean FoodSustainable aquaculture, mariculture, and regenerative seafood systems
Coastal Resilience & AdaptationFlood protection, living shorelines, and climate-adaptation infrastructure
Ocean Carbon / Blue CarbonOcean-based carbon removal, measurement, and blue-carbon markets
Pollution, Plastics & Marine DebrisDetection, prevention, and removal of marine pollution and debris

Additional Research

Source records are often incomplete — accelerator cohort pages may list only a company name, and federal award records can carry outdated contact details. AI-assisted web research is used to fill these gaps for each company:

Manual Review

AI classification and enrichment are a first pass, not the final word. Spot checks were conducted across the dataset to verify the accuracy of company descriptions and the validity of website links, with corrections applied directly to the affected records.

Every company in the dataset is also reviewed to verify ocean relevance, using the company website and other publicly available information. Companies appearing in multiple sources — a Seaworthy Collective graduate that later wins a federal award, for example — are matched to a single master record rather than counted twice. Misclassifications identified in review are corrected in the dataset, and review notes are retained alongside each record.

Data Updates

Sources are refreshed on a periodic basis: SBIR.gov quarterly as new awards are published, and organization rosters as new cohorts and award cycles are announced. The current dataset includes companies identified through July 2026. Classification is re-run when new data is added to ensure consistency across the full dataset, the ocean-relevance review is repeated for newly added companies, and cross-source deduplication is re-applied before new records enter the master dataset.

Limitations

  1. Source-text classification. Classification depends on the source text available for each company — award abstracts, program descriptions, and websites vary widely in depth. Where text is vague or generic, sector and theme assignment relies on analyst judgment and is flagged accordingly.
  2. Category boundaries. Some technologies span multiple sectors; tags reflect the primary focus of each company. Ocean-adjacent work by companies whose core business is elsewhere may sit outside the dataset.
  3. Discovery-channel coverage. Companies enter the dataset through federal awards and accelerator or award programs. Ocean-economy companies without those touchpoints — established operators, ports, tourism businesses, bootstrapped firms — are underrepresented, so the dataset reflects the emerging-innovation segment rather than the total ocean economy.
  4. Funding ≠ outcomes. Funding figures are amounts at time of award, not disbursements or outcomes. Funding is a leading indicator of innovation activity, not a measure of realized impact.
  5. Geographic assignment. Companies are mapped to their registered address, which may differ from where the work is performed.
  6. Small samples. With 131 companies across six sectors and seven themes, per-category counts are small; percentage comparisons between categories should be read with that base in mind.