This is the main documentation page for those looking to edit the Plant Humanities Knowledge Graph. Please look here for information associated with knowledge graph creation and updates about the Plants Humanities project.
In September 2018 Dumbarton Oaks and JSTOR Labs received partner grants from the National Endowment for the Humanities (NEH) to advance a Plant Humanities Initiative. These grants integrate elements of the digital humanities with scholarly programming, to set forth a new, interdisciplinary field that communicates the unparalleled significance of plants to human culture. The grant has three main goals: to provide innovative research and professional development opportunities for early-career humanists; to create a digital tool informed by the insights and needs of teachers and students as well as librarians and technical experts; and to supplement existing digitized resources with new primary source material, contextualize them, and disseminate them.
This knowledge graph aims to compliment that digital tool while extending current digitized resources. As such it employs Resource Description Framework (RDF) metadata standards that allow researchers to contribute to a semantic web architecture by cataloguing, navigating, and organizing data surrounding the cultural history of plants.
The JSTOR Labs Plant Humanities Knowledge Graph emulates the technical and philosophical architecture of Wikidata. It is a collaboratively edited knowledge base committed to producing linked open data under a public domain license.
Conceptually, this graph grew from the seed of Dumbarton Oaks' vast collection of herbals. While initial conversations centered on defining networks of catalogued plants, they rapidly expanded into exploring and recording the people involved in the creation and amelioration of these books. Not wanting to limit ourselves to traditional cataloguing ontologies, we have included ways of expressing this data beyond modern taxonomies. Where available, we include the ability to record things such as traditional uses, Galenic properties, and historic names. We additionally grant the ability to link to modern plant taxonomies. The benefit of a linked open data architecture is that these ontologies are not set in stone. Please let us know (via EMAIL?) if there are any other properties we should include.
The Plant Humanities Knowledge Graph is a document-oriented database focused on items, which represent topics, concepts, or objects. Each item is identified by a unique, "PID," number that enables basic information about an item to be stored without favoring any particular language. Fundamentally, an item consists of a label, a description, and some number of statements.
Editing and Contributing
This will include information about editing or creating an individual Knowledge Graph page.
Individual Knowledge graph entries can be updated or created manually once a user name has been created.
To add individual items:
- Click on the Add New Item tag on the left tool bar of this page.
- Enter a label, description, and alias (if applicable). Please note that aliases are separated by the pipe symbol ( | ). Click create.
- Once the base document has been created, feel free to enter any additional properties through the 'add statement' button.
- A master list of properties can be found here.
- It may help to see some examples of items with many properties. Find some useful examples below:
- Due to the experimental nature of this early-stage product, known data caveats are as follows:
- We only support images added from the Bioheritage Diversity Library. Thus, if you would like to link to individual pages within a book or herbal, please locate it on the BHL website and add the DOI under the property tag: full work available at (P134).
- If looking to add a linked property, for example, mentions in JSTOR Global Plants ... should we encourage this to be done manually?
This will include information for the fellows about updating the Knowledge Graph via worksheets.
Will hold off on updating until this workflow is in place. Keep this as two steps. One record mentions and the other records things (tested items).
- Find item on Wikidata. If this item does not exist on Wikidata, skip to step two.
- Copy Wikidata QID. For example, copy (Q14560).
- Paste QID into 'wikidata id column. Note that a user must manually paste in a 'wd:' A prefix is required. (wd:Q14560).
- If the wikidata id column is populated, the remaining fields will be auto-populated based on Wikidata's fields. If more information is desired, proceed to step two.
- Populate label, description, aliases, instance of, subclass of fields with string fields.
- Again, these fields will be auto-populated if the wikidata id field is filled.
- Once all information is populated, populate the 'Ingest Data' column with TRUE and click the 'Ingest Data' button
- NB: A green pop-up will appear on Google Sheets, informing the user that a script is running and that a script has finished, however, the data fields will still remain empty. This is a known data lag, please be patient as the database.
- Once the script has successfully run, a local id will be produced (Q1024 -- Need to change for production).
Conceptually, populating the mentions column operates in much the same way as items, however, certain caveats exist. Mentions exist as part of a singular item, namely an herbal or a book. They are plants on a single page and exist as entities with properties at a point in time.
Please see the steps below for more information:
- First identify the plant that is mentioned in the book or herbal.
- Ensure the book is loaded into our local knowledge graph. If not proceed to our guide on loading books into our knowledge graph.
- Identify the columns paste the following:
- The page number (on the physical book copy)
- The URL for the page.
- Please note that at this time we only accept Biodiversity Heritage Library URLs
- The historic name
- Multiple historic names can be added in the same cell with space or pipe delimiters, or in separate rows (preferred entry method).
- A language tag is required. A comprehensive list of language codes recognized by Wikidata can be found here.
- A local QID for the main subject or modern name for the plant (species or genus) EXAMPLE
- Plant Description is a free-form text field
- Has Virtue requires items from a controlled vocabulary of virtues. See list [NEED_TO_MAKE_THIS_PAGE here].
- Select one of the galenic qualities. This is controlled vocabulary representing the galenic properties and humors. See a full list [NEED_TO_MAKE_THIS_PAGE here].
- Plant origin string is a free-form text field.
- If possible, the images are demarked through the image: page, Image: creator, Image creator name string, and image: inception fields. Here, the user identifies the page range of the mention (usually one page). They also denote the engraver, or artist as identified by either QID or string. Finally, inception is used to identify the creation date of those images.
Knowledge Graph Query Service
The JSTOR Labs Knowledge Graph Query Service employs the standard SPARQL language to retrieve and manipulate data stored in the RDF format. SPARQL allows users to write queries as triples, conjunctions, disjunctions, and optional patterns. For more information on querying the Plant Humanities Knowledge Graph along with sample SPARQL queries please see our Plant Humanities Knowledge Graph Wiki.
Need to fill this in.