Many years of working with organisations to produce open data has consistently reaffirmed one thing: it’s important for people to preview their data before publishing it.

Preview both provides a way for people to check that they’ve put their data together correctly, and provides an immediate dose of motivation — it’s likely that they’ve never seen that data presented visually before. Publishing and maintaining data in the open is often a big step for organisations, so being able to effectively see and understand what any dataset might look like is an essential step — particularly when it’s an effort involving many different people.

Yes, we know people could check the spreadsheet, XML and/or JSON data directly — but many of the teams we’ve spoken to don’t see that as the ideal way to preview data. In Open Contracting, we’ve seen tools such as OCDS Show fill this need.

When working on data for the International Aid Transparency Initiative (IATI), an important tool has become the d-portal platform — a long-standing website that consumes and visualises all published IATI data, and updates every night. Over the several years of its existence it’s grown and developed, based on feedback from users such as ourselves. Visualising IATI data can be hard, given the multitude of data fields, specialist terminology and varied approaches to publishing — so it’s a real achievement that d-portal can act as a daily mirror up to the published IATI data.

In our work to support organisations understand and publish IATI data, we’ve also been able to generate “previews” of d-portal. These are versions of the website that are not live or accessible to others. Through this, organisations have been able to see how their data might look to the world tomorrow, should they decide to publish.

This preview function was originally shared with us directly as a “super secret link” — a “private mode” d-portal that creates copies of the site, with non-published data. Over time, we shared this secret link with others, and onwards — so the reality has become that the preview aspect of d-portal has become just as widespread as the live version!

We thought this wasn’t helpful or sustainable, going forward. So, we proposed to our cooperative colleagues that we would provide a budget to put this onto a secure public footing — enough for development work to take place, and then for an initial six months of hosting, to ensure it could be robustly used.

With the consent of the co-operative, we got to work. The team at Wetgenes (who maintain the core d-portal codebase) did some much-valued work to create a service that could generate d-portal previews through a variety of data input methods. After testing, they then built a simple user interface, and we also hooked our data validator and explorer tool CoVE into the mix, to enable users to generate d-previews as part of their data conversion workflow.

We then made some changes to the preview template, to make it much clearer to users that this is a temporary instance, and not live data nor official IATI infrastructure. We know that people can share these preview sites, in preparing to publish — so we wanted to support that, but also make it clear that these previews are throwaway.

With that in place — we’re pleased to announce d-preview. This is a service open to anyone — either through the upload user interface or programmatically.

We’re happy to cover the server and maintenance costs for the next six months, and will share how this is looking over the coming weeks. We’re very proud to be able to invest in this infrastructure, and delighted to position this on the Code For IATI domain.

Good luck with your previews — let us know how it goes!