Biodiversa

A public research portal for the Conservation Ecology Group, where every species page is assembled from external biodiversity databases rather than being written by hand.

 

## The client

Biodiversa is the website of the Conservation Ecology Group, led by Prof. Markus Franzén: a public research portal for field researchers and the general public. Its species information comes from the world's biodiversity databases, so the group's own writing can concentrate on its research.

It was delivered together with the [Quercus Project](/projects/quercus-project) and [Museo Avellonia](/projects/museo-avellonia) sites for the same client.



 

 





 

## At a glance

Site[biodiversa.se](https://www.biodiversa.se)ClientConservation Ecology GroupLiveSince July 2026PlatformDrupal 11AccessibilityWCAG 2.1 AAMADDev products[Species External Entity](/products/species-ext), [Editorial Intelligence](/products/editorial-intelligence), [External Entities](/products/external-entities)

 

 





 

## What we built

- Species pages assembled from Catalogue of Life, Wikipedia, GBIF, iNaturalist and the IUCN Red List
- A source note on every field, naming the database it came from
- GBIF occurrence maps and phenology, population-trend and recent-observation charts on every species page, built with Leaflet and Chart.js
- An Editorial Intelligence queue that reviews incoming external news for relevance to the site's own content
- A publications section with profiles for 308 external authors
- Full WCAG 2.1 AA accessibility and schema.org structured data, built into the theme



 

 





 

## Assembling a species page

Each field on a species page goes through a weighted chain of source plugins run by [Species External Entity](/products/species-ext), Catalogue of Life first, and takes the first one in the chain that returns a value. Catalogue of Life itself covers a little over 80% of species, so a real share of fields comes from another authority further down the chain, by design rather than as a last resort. Every field carries a note naming the database it came from.

Caching is deliberately uneven: each source plugin keeps its raw response for 7 days, but the page's assembled value is cached for only 1 hour, 5 minutes if the previous attempt missed. The gap is intentional, so a short outage at an upstream database cannot make content that was already on the page disappear.



 

 





 

## Editorial Intelligence

[Editorial Intelligence](/products/editorial-intelligence) is an editors' queue that scores incoming external news articles for relevance against Biodiversa's own 39 news articles, using on-site embeddings rather than keyword matching. For articles worth featuring, an editor can generate an AI tagline through the Claude API: 19 articles carry one so far.



 

 





## Start a project.

Tell us about your research group and what its portal has to do.



[Get in touch →](/contact)