🌊 OceanScope — Global Fishing Intelligence Platform

The Problem

The world's oceans generate enormous amounts of vessel activity data through AIS (Automatic Identification System). However, turning this data into useful intelligence is difficult.

Raw fishing-effort datasets can contain millions of spatial and temporal records. Simply displaying those records does not answer the questions that matter:

Where is fishing activity concentrated?
Which countries contribute the most fishing effort?
Which fishing methods dominate?
How does fishing activity change over time?
Which vessels and fleets are most active?
Where are unusual or near-continuous activity patterns occurring?
Which areas experience the greatest fishing pressure?

We wanted to build a system that could move beyond static charts and transform large-scale fishing data into an interactive intelligence platform.

💡 Our Solution

OceanScope is a global fishing intelligence platform built to explore, analyze, and interpret large-scale AIS-derived fishing activity.

Instead of forcing users to work directly with massive datasets, OceanScope provides a unified analytical interface covering:

🌍 Global spatial activity 🚩 Flag-state intelligence ⚙️ Gear-type analytics 🚢 Vessel intelligence 📅 Temporal trends 🗺️ Spatial density analysis ⚠️ Anomaly detection 🌱 Sustainability indicators 🤖 AI-generated insights 📦 Filtered data exports

The goal is simple:

Turn complex fishing activity data into intelligence that people can actually explore and understand.

📊 The Data

OceanScope uses the Global Fishing Watch v3.0 fishing-effort corpus for 2019.

The dataset contains large-scale fishing activity derived from AIS data and allows us to analyze fishing effort across geographic cells, dates, flag states, and gear classifications.

Our 2019 analysis contains:

52.1M fishing vessel-hours
162.9M AIS vessel-presence hours
32.0% fishing ratio
249 active flag states
16 gear classes
1.9M active 0.1° cells
68.5K registered vessels

This scale is what motivated us to design OceanScope around analytical aggregation rather than repeatedly processing raw records in the UI.

🔎 What OceanScope Can Do

1. Executive Overview

The Executive Overview provides an immediate global picture of fishing activity.

Users can see total fishing effort, vessel presence, fishing ratio, active countries, gear classes, active spatial cells, peak months, and busiest days.

The dashboard also automatically generates observations from the current dataset selection.

2. Global Fishing Map

The Global Fishing Map visualizes fishing activity across the world's oceans.

Users can explore fishing effort at 0.1° spatial resolution, zoom into hotspots, and apply filters for:

Date
Flag state
Gear type
Ocean region
Minimum fishing hours
Vessel presence

This allows users to move from a global overview to specific fishing areas.

3. Country Analysis

Country Analysis examines fishing activity by flag state.

It provides:

Flag-state rankings
Global effort share
Choropleth visualization
Flag → gear relationships
Country comparisons
Quarterly momentum

For example, our 2019 analysis shows that China accounts for 28.3% of global fishing effort, while the top five flag states account for 53.6%.

4. Gear Type Analytics

Fishing activity can also be analyzed by fishing method.

OceanScope supports 16 GFW gear classes and provides:

Gear hierarchy
Gear share
Flag composition within gears
Monthly gear trends
Daily effort distributions

Trawlers represent approximately 47.4% of recorded fishing effort in the 2019 corpus.

5. Time Trends

The Time Trends module analyzes how fishing activity changes over time.

It includes:

Daily fishing effort
Moving averages
Weekday patterns
Year heatmaps
Monthly changes
Seasonal analysis
Exploratory forecasting

This makes it possible to identify peaks, seasonal patterns, and changes in fishing intensity.

🚢 Vessel Intelligence

OceanScope connects fishing effort with vessel-level characteristics from the 2019 vessel register.

The platform analyzes:

Vessel counts
Flag distribution
Vessel classes
Vessel length
Engine power
Fishing utilization

The register contains approximately 68.5K unique vessels, with 87.6% reporting a fishing vessel type.

This helps bridge the gap between aggregated fishing activity and fleet-level characteristics.

🗺️ Spatial Analytics

The Spatial Analytics module transforms individual spatial observations into density patterns.

Using the 0.1° grid, OceanScope can identify:

Dense fishing areas
Fishing grounds
Spatial hotspots
Concentrated activity
Regional patterns

This provides a higher-level view of where fishing activity is persistently concentrated.

⚠️ Anomaly Detection

One of OceanScope's key intelligence features is anomaly detection.

We combine:

Isolation Forest
Statistical z-scores
Multiple activity features

to generate an anomaly index and surface unusual spatial or temporal behavior.

The system can identify areas with unusually high activity and cells exhibiting near-continuous fishing ratios.

These results are intended as analytical signals for further investigation, not as automatic claims of illegal fishing.

🌱 Sustainability Analytics

OceanScope also explores the environmental footprint of fishing activity.

The Sustainability module provides:

Fishing pressure per km²
Pressure distributions
Fleet exploitation indicators
Flag-level comparisons
Fleet profiles

The exploitation metric is a composite analytical indicator combining factors such as fishing ratio, effort per vessel, gear concentration, and seasonality.

This provides a framework for comparing fishing intensity across fleets.

🤖 AI Insights

We wanted OceanScope to do more than display charts.

The AI Insights layer automatically generates observations from the computed aggregates.

Insights are categorized into areas such as:

Scale
Concentration
Seasonality
Spatial patterns
Fleet behavior
Anomalies
Data quality

The platform also provides a natural-language interface where users can ask questions such as:

"Which country contributes the most fishing effort?"

"What is the dominant fishing gear?"

"When was fishing activity highest?"

Instead of manually searching through multiple charts, users can ask the data directly.

⚙️ How We Built It

OceanScope was designed around a high-performance analytical pipeline.

Data Layer

Global Fishing Watch v3.0

↓

Processing

Polars

↓

Analytical Engine

DuckDB

↓

Application

Streamlit

↓

Visualization

Plotly + deck.gl

↓

Intelligence

Statistical analysis + Isolation Forest + derived analytical indicators + natural-language insights

DuckDB allows analytical queries to be performed efficiently, while Polars provides high-performance dataframe processing.

Plotly powers interactive analytical charts, while deck.gl is used for large-scale geospatial visualization.

Streamlit ties the analytical and visualization layers together into a single interactive platform.

🧠 Why This Approach?

A major challenge was scale.

Large AIS-derived datasets are not practical to repeatedly load and manipulate directly inside an interactive dashboard.

We therefore designed OceanScope around pre-computed and queryable aggregates.

This allows the application to:

Process large datasets efficiently.
Reduce repeated computation.
Apply filters dynamically.
Keep visualizations responsive.
Generate insights from the same underlying analytical layer.

The result is an interface that feels like an intelligence platform rather than a static dataset viewer.

🚀 What We Learned

Building OceanScope taught us that the difficult part of large-scale data applications is not simply creating charts.

The real challenge is designing the pipeline between:

Raw Data → Processing → Aggregation → Querying → Visualization → Intelligence

We learned how important data modeling, aggregation strategy, spatial resolution, query performance, and visualization design become when working with millions of records.

We also learned that analytical indicators need to be communicated carefully. Metrics such as anomaly and exploitation indices should be treated as signals for investigation rather than definitive conclusions.

🌍 Impact & Future Potential

OceanScope can evolve into a broader marine intelligence system.

Future versions could incorporate:

Multi-year fishing activity
Near-real-time AIS streams
Protected marine areas
Weather and ocean conditions
Vessel ownership information
Illegal, unreported, and unregulated fishing indicators
Cross-year fleet behavior
Advanced geospatial machine learning
Automated alerts for emerging anomalies
API access for researchers and organizations

The architecture is designed so that additional datasets and analytical modules can be integrated without rebuilding the entire platform.

🔗 Explore OceanScope

🏁 Final Thought

OceanScope started with a simple question:

What if massive global fishing data could be explored as easily as a modern analytics platform?

We built OceanScope to answer that question.

From raw AIS data to global fishing intelligence.

🌊 OceanScope

Built With

ais
artificial-intelligence
data-science
data-visualization
deck.gl
duckdb
geospatial
machine-learning
marine-analytics
plotly
polars
python
streamlit
sustainability

Updates

Sahas Nagar started this project — Aug 16, 2026 03:10 AM EDT

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.