EMODnet Vessel Density Maps (EU Waters)

Most traffic data describes ships. This one describes water. The EMODnet Vessel Density Maps (EU Waters) turn raw AIS messages into a 1x1 km intensity surface laid over every EU sea basin and some neighbouring waters, where each cell stores how many hours vessels spent inside it, per square kilometre, per month.

The product is built by the EMODnet Human Activities project - produced by Cogea for the European Commission's DG MARE - from satellite AIS purchased annually from Collecte Localisation Satellites (CLS), ORBCOMM and vesseltracker.com. It cuts two ways. By ship type: thirteen documented codes, from 1 Fishing and 8 Passenger to 9 Cargo and 10 Tanker, plus an All types aggregate. By period: monthly slices, yearly averages and seasonal composites - 56 gridded density products in total.

Inside Datadory's 15-record marine transportation slice it holds a unique seat: the only gridded-density product in the shelf. Every other record ships points, tables or live messages; this is the one that answers where before anyone asks which ship. Get a sample of this dataset cut to your basins and ship types.

One observation per grid cell, per ship type, per period - the shape below follows the verified field dictionary:

# One observation per 1x1 km cell, per ship type, per period
grid_cell       : <1x1 km raster cell>   # EU sea basins + neighbouring waters
density         : hours / km^2 / month   # cumulative vessel presence
ship_type_code  : 9                      # 9 = Cargo; codes 0-12 plus All types
month_layer     : vesseldensity_01       # February slice; _avg and _seasonal exist
time_period     : 2017 ... 2024          # verified span, referenced by year and month

An honesty note first: the record ships no sample file, so the block above shows structure rather than pasted extracts - per-cell numbers arrive with your sample request instead of being invented here. Three things worth noticing anyway. The density column is an intensity measure, not a count, so a lane crossed in minutes reads far below a fleet working a box for days. The ship-type selector lives in the product variant itself, which makes “cargo only, February” a layer choice rather than a filter you have to build. And the grid is regular, which means the slices stack: twelve monthly layers make a season, stacked years make a trend, all on identical cells with zero resampling.

Four documented attributes carry the model, and the naming convention carries the rest. density is the payload - hours per square kilometre per month. ship_type_code selects the layer out of the thirteen-type scheme. month_layer distinguishes monthly slices from yearly averages and seasonal composites. time_period anchors the raster to its year and month. Definitions come from the publisher's own documentation and were verified during research, so the dictionary below is shape-exact rather than inferred; per-column naming for your pipeline is locked when your sample is prepared.

• Geography: all EU sea basins plus some neighbouring waters, on a uniform 1x1 km grid - the Baltic to the Black Sea, the Atlantic approaches to the Mediterranean, on one cell size.
• Temporal: 2017 through 2024, with monthly layers, yearly averages and seasonal composites stacked on the same span. The 2024 vintage carries a published caveat - satellite input loss from June onward that the provider attempted to mitigate.
• Granularity: one 1x1 km cell, per ship type, per month. Sub-cell behaviour stops at the grid boundary by design, because aggregation is what protects commercially sensitive movement.

On Datadory's 0-10 quality rubric the record scores 8, against a 7.81 average across the 1,744 cataloged datasets - verified field documentation and Europe-wide breadth at kilometre grain, held back mainly by the absence of shipped sample files and that 2024 input caveat.

API, files, or your warehouse. Daily, weekly, or hourly.

• Route benchmarking and network planning - compare carrier lanes across basins and years on identical cells instead of rebuilding intensity surfaces from raw positions.
• Port-call and congestion context - put arrival numbers against the approach corridors actually feeding them, by month, before promising dwell improvements.
• Offshore siting and cable routing - screen wind-farm plots and interconnector routes against fishing, tanker and cargo pressure before the first survey vessel sails.
• Environmental and regulatory analysis - separate fishing effort from transit traffic in protected areas, with a citable EU-produced evidence base.
• Market sizing by basin - read activity intensity across the Mediterranean or North Sea as a demand proxy that national statistics never aggregate.
• Cited reference work - anchor journalism and academic papers to a gridded EU product whose provenance runs to named AIS providers.

Data scientists and ML engineers rank first: a regular 1x1 km grid with a documented unit loads straight into raster pipelines with no track-replay preprocessing, and the cells stay identical across years (their marine-transportation workflows). Developers building data products get map-ready grids that render as heat layers without a spatial backend (builders' use cases). Market researchers and consultants read basin-level activity patterns as demand signals (their use cases). Competitive-intel teams benchmark carrier lanes quarter over quarter (their use cases). E-commerce operators place freight-exposure context around their supply chains (their workflows), and journalists, academics and students cite an EU-produced product with provenance down to the AIS providers (their use cases).

• MarineCadastre AIS Vessel Traffic Data (USCG Nationwide AIS) - the point-level counterpart for US waters: tens of billions of position reports from 2009 through 2025, where this grid trades per-vessel detail for Europe-wide intensity (MarineCadastre).
• Eurostat Maritime Transport Statistics (mar_go / mar_pa / mar_tf) - EU port gross-weight and TEU series counting what was handled, where the grid measures what moved past it (Eurostat).
• MarineTraffic Vessel Directory and AIS Data - the per-vessel lens: directories and particulars for when a question needs the ship, not the water (MarineTraffic).
• vs Data.gov Maritime Collection - the curated single product against the US federal aggregator of roughly 240 maritime datasets; field-by-field breakdown in the dedicated comparison.
• Glossary notes - what an AIS position report contributes before aggregation, and why dark vessel behaviour never reaches a density grid at all.

Two honesty notes before you request a sample. Aggregation protects privacy and commercially sensitive information, so quality varies and no per-vessel detail survives into the cells - pair with a point-based feed when a specific ship matters. And the 2024 vintage reads against a known satellite-input gap from June onward, so confirm whether a later release restates it before drawing trend conclusions. Full slice ranking sits on the marine transportation data hub, with the ranked view at best marine-transportation datasets.