Biodiversity's Heritage Library's archives represent an enormous collection of digitized books from the 15th-21st centuries. It is the world's largest open access digital library for biodiversity literature, hosting over 59 millions pages from hundreds of thousands of volumes. This repository presents an attempt to valorize this priceless heritage using computational methods and dynamic visualizations. A 3D tree of life, with leaves blooming into the stunning scientific illustrations contained in these archives, is developped to allow anyone to wander freely through (a small sample of) this database.
The dataset used here was built upon a sample of the illustrations posted by BHL on Flickr — to enrich pictural content with metadata (specie taxonomy, artist informations, ...) through crowdsourcing — via Flickr's API. The content was then augmented with taxonomic informations thanks to NCBI database which allowed to retrieve all 6 main taxonomic ranks from the specie level (harvested from Flickr's machine tags) up to the domain (Eukarya).
This repository hosts two independent visualizations, developped as a semester project for the Cultural Data Sculpting class of EPFL (Lausanne, Switzerland). Both of them rely on the Force-Directed Graph library developped by Vasco Asturiano (vasturiano).
The first one: KODAMA, is a 3D-tree of life whose leaves bloom into the collected illustrations archived by the BHL when clicked. It is built with 3D Force-Directed Graph, which uses d3-force-3d for the physics of the graph and ThreeJS to render it. Possible interactions with the components of the tree are the following:
- left-click on node: focus on node
- left-click on leaf: display all illustrations that share the same taxonomic class as the clicked leaf
- right-click on leaf: display caption below the illustration (name of the artist, title of the book containing the illustration, year of publication of the book) — caveat: book title and date of publication have been harvested from Flickr's descriptions and might not be totally accurate
- cmd+click or ctrl+click on leaf: open in new window the page of the illustration within the book it originally appeared, on the archived volume hosted on BHL website (please use this command if you want more correct and more complete informations than displayed in the caption)
- left-click on link: focus on target node
The second one: KODAM.AR is a coarse extension of the first tree into an AR module. It was implemented thanks to 3D Force-Directed Graph in AR which makes use of AR.js with A-frame for rendering. To load this AR extension, make sure that the camera of the device on which you want to test it points toward the image of the marker displayed below (you can also download it from here):
Use the following links to explore these propositions of visualization of BHL's archives:
Disclairmer: the visualization window might not show the full canvas.