Improve Document Flexibility

[emily-noble-SSP]

Knowledge forces any documents into the Document entity type; however it may be more appropriate to model the documents as multiple entity types. For example, a construction company may have multiple documents for a project (each with their own distinct attributes): bids, closing documents, permits, designs, ROW references, parcel scans, status reports, etc. Grouping all of these documents into a single entity type may limit (or complicate) analysis of a network. Another way to think of this is allowing documents to be added as properties of an entity, rather than entities themselves (in the same way shape is a property of an entity, rather than a separate entity). 

[AdamMartin]

Thanks Emily for your feedback. I agree, Documents could be represented as multiple types of Entities, or as you are describing, could be represented as properties on other Entity types. 

Another pattern for using documents would be as Provenance records (i.e. source documents) related to specific properties on your entities.

To clarify, Documents are a special default type of entity when you create your Knowledge Graph. However, we are very data model agnostic, in that there's nothing in the software preventing you from modeling multiple Entity types - like "Bid Document" "Closing Documents" etc that actually represent different types of documents, where those Entity types have similar properties like Document filepath URL, Title, etc.. Having said that, some of the current special functionality in the user interface specific to the default "Document" entity type would not be replicated into these other Entity types... (e.g. drag & drop, stripping all text from the document into a searchable property field).  One key workaround would be to have a Document type, and then a property such as "DocumentType" where you can specify "bids, closing docs, permits..." 

We've also seen graph modeling scenarios where it may be best to create Document entities, but also break out individual pages (Pages) or chunks of pages into their own Entities and use those to relate to specific other assets, land parcels, project entities in your graph.

So am I correct to summarize your feature idea something like: Have option to apply "Document" entity type features to other user-defined entity types?

Are there other specific aspects you'd want us to capture as feedback when thinking about enhancements?

Best,

Adam