The Art of Life project

Editor's Notes

• #4: In 2006 libraries from botanical garden and natural history museums in the US and UK came together and decided they wanted to digitize all of the public domain literature in their collections and place them online to make them more widely accessible.One of our primary audiences are taxonomists who use BHL to find the first occurrence of name for a species in the historic literature. Also to track how that name has changed over time.
• #5: Today there are 15 institutions who are BHL members and who contribute content to the repository. As you can see from the list its made up of some of the largest and most well known nat history and botanical garden libraries and as well as some libraries who are much broader in scope but contribute their biodiversity specific materials to the repositoryMost recent member to join is the Library of Congress
• #6: The BHL is made up of 6 fulltime staff located at both the Smithsonian Libraries and Missouri Botanical Garden. Program Director, Martin Kalfatovic; Program Manager, Grace Costantino; and Collections Coordinator, Bianca Crowley are all based at the Smithsonian LibrariesTechnical Director, William Ulate; Programmer, Mike Lichtenberg; Data Analyst, Trish Rose-Sandler based at the Missouri Botanical GardenWe also have contributions from staff at the member institutions who allow a certain percentage of their staff time to work on BHL (when we tried to quantify how much time is spent by the part time staff it comes out to a little over 16 FTEs from the member institutions)
• #7: What began as a consortium between libraries in the US and UK it is now increasingly a global effort – there are BHL nodes in China, Australia, Egypt, Europe, Brazil and soon to be Africa! William also heads up our global coordination efforts. Each global partner maintains its own portal that is specialized to the needs of its users but we work to share content and technology across nodes and also try to not duplicate digitization efforts so we can maximize limited scanning funding.
• #8: The url for the portal is at biodiversitylibrary.orgHere is a sneak peak of our new UI that will go live on Monday. This new UI is a collaboration between BHL US and BHL Australia who have been maintaining separate portals but decide that based on feedback from BHL users that we should merge the 2 sites. We did a usability study in 2011 of the 2 sites and feedback indicated that Users wanted the look at feel of the Australian site but with the functionality of the US/UK site.After 6 yrs of digitization we have a critical mass of contentWe have over 57 thousand titles108 thousand volumesAlmost 40 million pages of text
• #9: You can browse BHL by Title, Author, Date or Collection
• #10: In our advanced search you can search on the title of a book, journal or article. You can search on Authors, Subjects and Scientific Names.
• #11: Scientific names are a particularly critical access point for our taxonomists who need to verify the first mention of a species in the published literature as this will help them validate which is the accepted name. In this case searching on zea mays brings back all of the books and journals that have mentioned zea mays anywhere within the text. you can see it was first mentioned in 1797. Our scientific name searching is enabled by taking the OCR output from each page of digitized text and running it through TaxonFinder, a taxonomic name recognition algorithm that is maintained by uBio and which we can call up via web services
• #12: Probably the most significant change to our UI was to the book viewer. Users now have ability to scroll multiple pages at once and to go directly to sections of the book like chapters and articles. When viewing a single page in the viewer you can see all the scientific names found on that page. The OCR can also be viewed side by side with the image of the page.
• #13: Because one of the founding principles of BHL was to make biodiversity lit available for open access and responsible use as part of biodiversity global commons we seek to provide data in a variety of forms for other to harvest and re-use One of the keys to making data harvestable is clearly stating for your users your copyright and licensing terms are so that they know what they can and cannot do with your data. Because most of our books and journals are historic literature that were published between 1450s- and 1923 they fall within the public domain - users may reuse them for either commercial or non-commercial purposes w/o permission. We do have some copyrighted material that we’ve digitized with the permission of the publisher. Those we provide under a Creative Commons Attribution/Non-Commercial/Share Alike License Our metadata (which includes catalog records and scientific names ) are available under a CC0 1.0 Universal (CC0 1.0) Public Domain Dedication license. Which essentially says we are the creaters of the data and want to dedicate it to the public domain. This allows anyone to reuse, modify, repurpose, and distribute the metadata for all purposes including commercial and non-commercial, with no need to ask for permission. This follows in the footsteps of other libraries such as The British Library, Europeana, the University of Michigan Library, and Harvard who have adopted it for their online catalog data. 
• #14: Therefore we provide data in a variety of exportsWe encourage digital library aggregators to incorporate our records into their portals and library catalogs. We provide data in ways that it can be mined recontextualized
• #15: Besides the traditional harvesting of our metadata records into other portals such as OCLC WorldcatAnd Digital Public Library of America our content has been mined and recontextualized for a variety of interesting purposes.  The website and webservice BioStor by Rod Page provides tools for extracting, annotating, and visualising information on literature from BHL (http://biostor.org/). In this example, Rod has identified articles found in the Proceedings of the United States National Museum. We are now incorporating Rod’s articles back into the BHL portal and they will be searchable in the new UI. So here’s an example of how making your data open accessible could result in real benefits back to your institution and users
• #16: Ryan Schenk is using publication dates of works in BHL to build histograms of the number of publications-per-year for specific species, In this example, the Guniea Pig (http://synynyms.no.de/ )
• #17: Another key to harvestable is promotion. People won’t know your data is out there and can be mined unless you advertise and really get the word out. We also make extensive use of social media to let users know about new content, contextualize the existing content, interact with users and in general let users know we are a constantly dynamic and growing repository. We do this via: our Blog, Twitter, Facebook and Pinterest.
• #18: So all of this background information was provided in order to give context as to the need for the Art of Life project. Problem Statement- Art of Life evolved out of a need in the BHL that was expressed by our users. We had a critical mass of content online, BHL users knew there were amazing images within the BHL pages but there was no easy way to find them other than opening up a BHL book or volume and scrolling through page by page to find illustrations. There is no descriptive metadata attached to the illustration that would tell you the content of the image, date when they were created or who was involved in their creation.
• #19: One way we’ve tried to address this need is by pushing selected images to Flickr. We have created a BHL account in Flickr and pushed over 63,000 images so far but but this is all a very manual process that takes considerable staff time. We estimate that we have millions of illustrations within BHL so this manual process does not scale well
• #21: This is the Art of Life workflow diagram which identifies the 4 processes the illustrations will go through as they move through each stage of the workflow. They include: Extract, Classify, Describe, and Share.The Extract stage is where BHL pages will be run through an algorithm to identify which pages contain illustrations, whether they be full plates or only a section of the page. This algorithm is being developed by our partners on the project from the Indianapolis Museum of Art Lab. At the Classify stage, the pages with illustrations will be tagged by Art of Life staff as being one or several broad types such as drawing/painting, photograph, diagram and even map. The tool we will use for this is actually an existing tool created at the Smithsonian Libraries called Macaw that BHL uses to for books scanned outside of its traditional workflow.For the Describe stage, the illustrations will be pushed into platforms such as Flickr and Wikimedia Commons where both the general public and specialists can describe them in much greater detail such as adding a title, creator, date (if different from date of publication), and subjects. Wikimedia Commons is where the schema can play a role. Because Wikimedia allows you to create templates we can provide guidance to taggers on what information to record and how to record it. In the Share stage, the metadata for the illustrations will be reingested back into the BHL portal for searching there. And of course we want to be able to preserve any contributed metadata from external platforms. We also want to broaden the audience for these illustrations because we believe they have a wide appeal to artists, biologists, humanities scholars, particularly historians of science; librarians, education and outreach. Many of the audiences don’t know about BHL and won’t go to the BHL platform looking for the content so we want to push the illustrations out to environments where they already are: Encyclopedia of Life, ARTstor, and even ITunes.
• #24: IMA developed a UI for us to review the results of each algorithm. This screen shot just shows results from the OCR files (ABBY) and contrast. Those were the 2 algorithms we found to be most accurate. Its showing you all the pages in a book, how many actual illustrations and how many true/false postitives. The actual illustrations is based on a gold standard testset that we developed where we’ve manually gone and identified which pages have images so the algorithm can always compare against the test set and come up with a percentage of accuracy.
• #26: We ended up choosing most of the schema elements from the VRA Core because its elements and attributes were mostly closely aligned with the types of information we felt were important to record. But also because its relationship of works linking to one or more images fit nicely with the BHL pages often containing one or more illustrations on a single page. The only thing the VRA Core lacked was a way to record an acceptedName and CommonName for a species. VRA Core has a subject attribute type of scientificName but Taxonomists would be interested in knowing the multiple names by which species are known. Darwin Core was able to fulfill this need and so we borrowed 2 elements from that schema.
• #27: Here are the elements chosen (this is still in draft form)
• #28: Here is an illustration described using the schema