Knowledge Maps let you build semantic fields or ontologies of terms, which you can then use to label resources you create in Mandala. At the most basic level, Knowledge Maps are terms or tags in a hierarchy. This places a term in the appropriate context. Furthermore, each Knowledge Map term has its own associated metadata, including term definitions, related terms, alternate names, specific GIS locations, and more. You can see all this information using the Knowledge Maps viewers. Whenever you label a resource in Mandala with a Knowledge Map, it links to the Knowledge Map viewer. There are currently two types of Knowledge Maps: subjects, terms, and places. Each has its own viewer at mandala.shanti.virginia.edu/subjects and mandala.shanti.virginia.edu/places, respectively. We call the hierarchical grouping of terms a Knowledge Map tree.
Subjects, terms, and places are built collaboratively, through crowd-sourcing. Ideally, new subjects or places should integrate into the existing trees at https://mandala.shanti.virginia.edu/subjects and https://mandala.shanti.virginia.edu/places. We understand that everyone has different needs for their project, and are happy to implement whatever terms most serve your scholarship. We only ask that you come to meetings open to collaboration and productive conversation. The first step in adding new Knowledge Maps terms is to email the Mandala team at mandala@virginia.edu. We'll set up a meeting to discuss your project. |
This guide helps you prepare for your first Knowledge Maps discussion.
Before Your Meeting
Develop captions alongside subject names . Captions are brief summaries of a subject under 140 characters. This helps you justify a subject's place in the tree. You can reuse them when you move into Mandala.
Refer to experts and existing literature to improve your tree. You also don't always need to start from scratch. Ontologies are ongoing fields of research, so there may be an existing ontology that fits your needs. You can talk to us about how you can reuse these existing ontologies in Places at mandala@virginia.edu .
|
|
---|---|
|
|
Make sure the group of subject in a single level is cohesive. This means every subject in that group has a similar "type" and "scope." Don't put very specific subjects alongside general subjects.
Make sure there is a clear flow for each path in the tree. There should be a logical progression from the topmost point of the tree to each individual subject at the bottom.
Make sure the flow goes from broadest to narrowest concept.
Avoid repetitive tree structures. Remember, you can use two tags on an asset to suggest a relationship that isn't represented in the ontology. Using the example below, you could tag assets with both the Education Stage and the Interventions.
|
|
---|---|
|
|
Use the places viewer (by searching or browsing ) to determine whether an existing place suits your needs.
This can be tricky to decide, especially if there's a place that seems close to what you need. There aren't any "right" or "wrong" reasons to create a place, and these decisions will depend on your project. But here are some things to consider:
Example: Los Angeles renames the street Rodeo Road to Obama Boulevard.
In Mandala Places: You can make 'Rodeo Road' an alternate name, and set 'Obama Boulevard' as the new primary name. You can also add a note describing the change. You wouldn't need two separate places. Learn more about Alternate Names.
Example:
The Powerhouse Museum in Sidney, Australia decides to move to Parramatta, Australia. The organization stays the same.
In Mandala Places:
You can change the coordinates from Sidney to Parramatta under 'Locations,' then add a note about the change. You wouldn't need two separate places. Learn more about Locations.
Example:
A monastery becomes a granary for ten years during the Cultural Revolution.
In Mandala Places:
You can track this change in the 'Feature Types' section, where both "granary" and "monastery" can be used. You can also add dates for each feature type. You wouldn't need two separate places. Learn more about Feature Types.
Perspectives
Formatting Trees in a Spreadsheet
Knowledge Map hierarchies in Mandala are built with tree data structures. Each knowledge map term is a node on the tree. You can place a knowledge map term using its parent — that is, the higher-level term the knowledge map falls under. Nodes that fall under a parent are its children. If a node doesn’t have a parent — that is, it’s the highest possible level of the tree – we call it the root node. A node may have many children, but it only has one the parent. In the example above:
To format your tree in a spreadsheet, every row represents a node. In the first column, write the node's name. In the second, write the name of the node's parent. If the node is a root, the parent will be blank. Make sure you use consistent names, with consistent capitalization. Here's a table for the sample tree:
Node | Parent |
---|---|
Architecture | |
Building dimensions | Architecture |
Building materials | Architecture |
Building layout | Building dimension |
Courtyard structure | Building layout |
Mansion | Building layout |
Large Mansion | Mansion |
Massive structure | Building layout |
courtyards | Building dimensions |
Stories | Building dimensions |
Mud | Building materials |
Mudbrick | Building materials |
Stone | Building materials |
This spreadsheet will help developers import your terms quickly.
Example: Grand Budapest Hotel, Gotham City, River Styx
Use the name however it appears in the source for your place's name -- this can be oral or a literary source.
You can give it the same name as the feature type. For example, a building with an unknown name can simply be called 'Building.' You can use the Knowledge Map id and other attributes to differentiate places with the same name.
Once you have added a new set of terms to a tree, we hope you'll become the "owner" for that node. That means you can add and delete any children for which that node is a parent. As the node owner, you can let collaborators add children to that term.
The timeline will vary. Places may take longer to add than subjects, especially if you need accurate administrative data. In these cases, adding places may take as long as a month. In your meeting, the Knowledge Maps team can give you more accurate time estimates based on your specific needs. Needless to say, you should start the process well before any deadlines.