We’re about to move into the next phase of development of Kasabi, switching from the private beta that we launched at the end of March into a full public beta.
This morning we’ve rolled out a new openly accessible version of the site which is at beta.kasabi.com. As a key milestone for the public beta, we’ve added the ability to create and publish your own datasets. Any registered user can now create and publish new datasets and APIs. Signing up to use an API on any dataset is a really simple click-through process.
Here’s a quick run down of some of the key features of Kasabi that you can find in the public beta:
- Free RDF data hosting
- Free Linked Data publishing under data.kasabi.com
- Search and browse to find datasets of interest
- Ability to add and configure basic branding, overview documentation and developer guides for every published dataset
- Create and configure datasets and APIs in “draft” status before publishing them for wider use
- Machine-readable views of dataset metadata, including pointers to APIs to allow boot-strapping of clients
- Standard set of APIs for every dataset, including support for the Google Refine Reconcilation API
- Data management APIs for populating and managing a dataset, including support for importing RDFa into a dataset
- User created APIs via SPARQL Stored Procedures and the Linked Data API
- Experimental API Explorer to quickly allow you to get hands-on with the data
- Ability for users to document and share SPARQL queries for SPARQL endpoints, lowering barrier to entry
With the public beta we’ve also brought a number of new datasets to the Linked Data cloud. This includes data from English Heritage, the Prelinger Archives, the Foodista wiki and (my favourite!) the Bricklink Lego parts database!
Data hosting and publishing will be a free aspect of Kasabi. Our business model will be largely based around transactional charging for premium services and high-volume use of data. Our goal is to bring some sustainability to the publishing and use of both free and commercial datasets.
During the initial beta period, publishing is limited to a maximum of 5 datasets. This is to allow us to manage the expected demand. Initially the data stores are geared up towards supporting the publication and management of smaller scale datasets: right now we’re most interested in helping people publish new, live, regularly updated data. That’s where the most value is.
When a dataset is created the storage is automatically created. Alongside each set is also created a standard set of APIs including Search, SPARQL Endpoints, and Reconciliation services. We’ve got a comparison chart that shows the features of the different APIs. The API documentation includes more details on how to use each API, including the data loading functionality.
Every instance of every API also has it’s own homepage that includes a quick reference for developers. Here’s an example SPARQL endpoint for the Foodista dataset.
Linked Data publishing is a built-in feature of Kasabi. If you mint your URIs under data.kasabi.com then you will also get Linked Data “out of the box” for each of you published datasets. In future releases we’ll be supporting custom domain hosting, so you can use your own APIs, but the basic free service is ready for use now.
We’ve got a very busy roadmap sketched out for the coming months. There is much more that we want to do to help make data to be both easy to publish and easy to use. Improved data management tools and client libraries will be one aspect of that. We also have some really exciting plans for how to make it easy to aggregate and remix datasets. Of course, we’ll also eventually begin rolling out the commercial aspects of the site. But, right now, we want to get the site in front of as many people as possible and ensure that the tools and APIs work as well as they can.
We remain committed to continuing to support the publication of free open data, and to that end the Talis Connected Commons programme will eventually become part of Kasabi.
We’re indebted to our early private beta testers for some great feedback from our hack day and in subsequent conversations. We’ve made a number of improvements to the application based on their feedback. Inevitably there’s always more that can (and will) be done. Useful constructive feedback is fantastic. We look forward to hearing more of it as you start to use Kasabi. We’re excited to see what you can create!