The world we know is built on an intricate web of connections. Every ecosystem, community, and habitat depends on a constantly shifting network of relationships that can be difficult to see, even with the best available data.
Maps have always been an essential tool for understanding that world. Biodiversity maps can show where threatened species have been observed. Predictive habitat models can estimate where species are likely to occur in the future based on environmental conditions. These products play a critical role in conservation planning and scientific research.
But ecosystems are more than a collection of individual species occupying the same space.
A pollinator depends on flowering plants. A bird depends on insects and nesting habitat. A fungus may support the health of an entire forest through invisible underground relationships. These interactions create the ecological fabric that sustains life. When one thread is removed, the effects can spread far beyond a single species.
When you look at a map of local flora and fauna, can you immediately see those dependencies? Can you visualize the chain of relationships connecting a vulnerable pollinator to a flowering plant, or an endangered bird to the habitat it relies on?
That question became the inspiration for Extinction Explorer.
Extinction Explorer combines an ArcGIS Knowledge Graph, a 3D food web, and a 3D map of species observations to make ecological relationships visible. The application allows users to explore both where species exist and how they are connected, transforming a collection of observations into a living network that can be explored interactively.
Building the Graph
At the core of the application is an ArcGIS Knowledge Graph built from ecological observation and species relationship data.
The graph consists of three primary entities:
- Species: plants, animals, fungi, and other organisms observed within the study region.
- Observations: individual sighting records containing location and temporal information.
- Documents: source records that capture where observation data originated.
Species relationships are incredibly complex and constantly evolving. Fortunately, initiatives such as Global Biotic Interactions (GloBI) aggregate ecological interaction data from scientific literature and biodiversity research communities, helping reveal how species interact across ecosystems.
For this project, those interactions were normalized into eight primary relationship types: PREDATES, POLLINATES, HOST_OF, PARASITIZES, NESTS_IN, SYMBIOTIC_WITH, ASSOCIATED_WITH, CO_OCCURS_WITH
Together, these relationships form a network that can be queried, analyzed, and visualized within ArcGIS Knowledge Graph.
Why ArcGIS Knowledge Graph?
Traditional biodiversity applications are often built around layers, tables, and joins. They excel at answering questions such as:
- Where is this species found?
- How many observations exist?
- What is its conservation status?
However, answering questions about relationships quickly becomes difficult.
- Which pollinators support this plant?
- Which predators depend on this species?
- What other organisms are affected if a species disappears?
ArcGIS Knowledge Graph provides a natural framework for modeling those questions.
In Extinction Explorer, species become entities and ecological interactions become explicit relationships. Using openCypher queries, the application begins with species classified as Critically Endangered, Endangered, or Vulnerable, then expands outward through one, two, or three levels of ecological connections.
Those same graph queries also retrieve associated observation records, allowing a single data model to power both the network visualization and the map.
At the three-hop level, the San Bernardino National Forest graph contains roughly 3,000 species, 8,000 ecological relationships, and tens of thousands of observations.
Two Views, One Ecosystem
The application presents the same ecosystem through two synchronized views.
The Map view displays thousands of observations collected through community science initiatives such as iNaturalist and other publicly available biodiversity datasets from GBIF. Observations are rendered as 3D pushpins within an ArcGIS SceneView, enabling users to explore species distributions across the terrain of the San Bernardino National Forest.
The Network view visualizes those same species as a 3D force-directed graph.
Species become nodes.
Relationships become connections.
What was once a collection of points is transformed into a living network of ecological interactions.
Because both views are driven by the same ArcGIS Knowledge Graph, they remain synchronized. Selecting a species in one view immediately highlights it in the other.
Rather than switching between separate applications, users can move seamlessly between geography and relationships.
Making Ecological Change Visible
The most compelling part of Extinction Explorer begins when a user selects a species.
Each selection can be treated as an extinction event.
When a species is removed from the network, Extinction Explorer calculates how that loss affects neighboring species and visualizes the consequences across both views.
The application records these impacts in an Extinction Log, documenting:
- Severed ecological relationships
- Newly isolated species
- Cascading effects throughout the network
Species affected by the event are highlighted using distinct colors, allowing users to see disruption spread through the ecosystem.
As more extinction events occur, ripple effects become visible across both the network and the landscape.
This experience helps bridge the gap between ecological theory and intuitive understanding.
Instead of reading about cascading ecological impacts, users can watch them unfold.
Just as importantly, every action can be reversed. Species can be restored and ecological connections rebuilt, allowing users to explore both disturbance and recovery.
Under the Hood
From a technical perspective, Extinction Explorer was designed as a lightweight web application focused on rapid experimentation.
The mapping experience is powered by the ArcGIS Maps SDK for JavaScript, using a SceneView to render thousands of observations as 3D pushpins over terrain. Observation status is managed through a renderer driven by an alertStatus field, allowing extinction cascades to update only affected features instead of rebuilding entire layers.
The network visualization is built with 3d-force-graph, three.js, and d3-force-3d, creating an interactive 3D representation of the ecosystem. Threatened species are symbolized using IUCN-inspired colors, while broader taxonomic groups are assigned their own category palettes.
Behind both visualizations sits an ArcGIS Knowledge Graph queried through openCypher.
[ExtinctionExplorer] openCypher (seeds):
MATCH (s:Species)
WHERE toUpper(s.iucn_status) STARTS WITH 'CR'
OR toUpper(s.iucn_status) STARTS WITH 'EN'
OR toUpper(s.iucn_status) STARTS WITH 'VU'
RETURN collect(DISTINCT s) AS nodes
The application dynamically expands through ecological relationships, retrieves observation records, and constructs the connected network displayed in both views.
One of the more interesting design decisions was keeping the architecture intentionally simple. The application runs entirely as a client-side experience using modern ES modules and the ArcGIS Maps SDK's module-loading capabilities.
Connecting GIS and Relationships
The San Bernardino National Forest serves as the prototype dataset, but the underlying pattern is much broader.
Any ArcGIS Knowledge Graph that models entities, relationships, and observations could support a similar experience. Whether the subject is biodiversity, infrastructure, supply chains, emergency management, or utilities, the combination of spatial context and graph relationships opens the door to new ways of exploring complex systems.
Traditional GIS helps answer where things happen.
Knowledge graphs help answer how things are connected.
Extinction Explorer explores what happens when those two perspectives come together. By linking a map of real-world observations with a dynamic network of ecological relationships, the prototype transforms a collection of points into a living system that users can interact with, question, and explore.
Try the demo: Extinction Explorer
