Triple Builder — Subsets

The Tool: https://tb.semlab.io/

The fundamentals of Linked Data are pretty straightforward at a conceptual level. The difficultly comes in like most things with the details. What is the URI of this predicate? Can I say this literal is in another language? How does this triple look serialized? This is where the learning curve usually is, how do I actually make a RDF statement and what does it look like?

When I started teaching metadata and linked data I made little tools for my students to use that reduced the complexity of building RDF. The first one I made I simply called Triple Builder which allowed creating collaborative RDF statements.

Old Triple Builder

It was a good ice breaker to start talking about RDF and the minutia of its implementation. This tool is 6 or 7 years old and I wanted to refactor it but make it even more useful beyond a pedagogical tool.

Github has a tool call Gist which lets you make little snippets of code and share them. I wanted to do something similar but with RDF, which would allow you to quickly create some triples and share it with others. This isn’t to replace very rigorous applications like Protege but a way to try out some quick modeling and share simple snippets of RDF.

I also wanted to embrace the spreadsheet which is the workhorse of any metadata operation. It is a pretty common trope in the metadata world about spreadsheets being at the center of everything.

https://tb.semlab.io/

This first version of the new tool allows for the creation of RDF using a spreadsheet interface with autocomplete that is preloaded with many common vocabularies.

When you begin you get a unique URL, you can make future edits to the graph by saving this address. At the top of the page is a spreadsheet interface that lets you enter the subject, predicate, object and literal language or data type. It also can do most things you can do with modern spreadsheets like copy and paste cells and rows, duplicate a cells, etc.

Spreadsheet interface

When you create a triple you will need a subject, this is normally the thing you are talking about. If you are saying something about something famous it probably has a URI somewhere in a name authority like id.loc.gov or wikidata.org. If not you can simply invent a fake URI like “http://www.example.com/entity1” or if you want to be thorough we have a sand box Wikibase installation that you can mint a new URI for whatever you would like to make triples about.

Click the link to create a new entity in our Wikibase sand box to use its URI

The tool will serialize out your triples into N-Triples, Turtle and JSON-LD. If you would like to share your work with others you can click the share tab and pick a URL to use:

This allows you to share a read-only version of your subset with others while not having to share your editable link. On this share tab you can add a note or other textual info that will be displayed along with your triples. It should even make a little twitter card with an image of your graph if you post it on social media:

The share page looks like this: https://tb.semlab.io/share/mattfoaf

Share page

At the top is a “Fork Subset” button which allows anyone viewing your data to make an editable copy for them to modify.

The next release of the tool will enable collaborative RDF creation, the original purpose of the Triple Builder. But for now it is a simple tool for trying out some RDF modeling and sharing it with others.