What roles are required for an effective portal? And can portal design better support a diversity of stakeholders?
In this post, part of a series exploring the past, present and future of data portals, I want to turn to the people and processes behind the portals.
For most open data portals it’s relatively hard to get a sense of who sits behind the site. Is there a full-time team of portal maintainers? Or are you looking at an online platform that has a few hours maintenance a year?
How many people are involved in running a standard data portal? Where do they sit within the organisation? How aware is the rest of the organisation of the portals existence? I’ve found relatively little research that addresses these questions: potentially a space for a future survey or other data collection. However, there are a couple of different perspectives on the organisational role of portals that we can pick up from the literature and from the evolution of portals over the last decade.
Early portals were a political project. Politicians and policy makers, or at least the people briefing them, were aware of portals as new announcements to be made, and there was a degree of policy-competition between portal owners over the number of breadth of datasets on the portal at launch. In many cases this empowered the teams working on open data initiatives to use the portal as a tool for going into the organisation and encouraging opening of different datasets.
Getting a dataset listed on a portal is a measurable achievement: and by organising portals around agencies and departments, it became possible to track (albeit very imperfectly) which parts of the organisation were making progress on opening their data (by quantity at least).
Though metrics on how up-to-date data is, or how standardised and complete certain datasets are, portals today can still be used as management tools, and, with suitable redesign, may be able to support management of much wider data policy across an organisation. However, it’s not clear how many portals teams have retained (particularly a number of electoral cycles later) the kind of high-level backing that helped predominantly technical teams to secure time and attention from other parts of the organisation to focus on data reforms.
The other important question (inherent to the long-history of digitisation in government - see Agar (2003)) when it comes to the teams around data portals and open data initiatives is whether (open) data management is seen as a task to be managed centrally by a specialist team, or as a core skill and responsibility to be spread throughout the organisation. ‘Portal people’ can either come into departments to take ownership of the public-facing version of datasets, and to act as a buffer between the organisation and outside data uses, or to act as mentors, facilitators and support team for the flow of data (and conversations) from internal datasets (and their related organisational functions) to external stakeholders. The ODI’s Olivier Thereaux frames this in terms of a shift from ‘being a better data publisher’, to becoming ‘a trusted guide and steward to the data community’.
Eaves, Mcguire and Carson note  a shift for open data teams from managing portals, to becoming internal analytics providers that may speak to the future role of the portal as just one tool embedded within the work of a wider organisational data analytics function.
One of the most common complaints about portals is that of poor metadata quality. Both basic dataset descriptors, and richer contextual descriptions that would help responsible use of a dataset, can be weak. But what’s the human process by which this metadata is created? Does it involve filling in relatively featureless online forms? Or are there richer processes to encourage data owners to think through the information that an outside data user might need? As projects like Civic Software Foundations’ Data Feminism inspired ‘Library of Context’ prototypes explore, good metadata potentially takes significant time to create.
How can organisations make the time spent generating good metadata into a valuable and valued process? (And how can they prioritise, to invest in the metadata that matters most?)
Previous work at the Open Data Institute has mapped a number of the pain points and challenged faced by current data publishers, and pointed to a number of areas for both improved tools and processes to help overcome barriers to data availability.
When we came up with the Five Stars of Open Data Engagement back in 2012, the third star was:
★ ★ ★ Support conversation around data
With the follow up questions:
Can people comment on datasets, or create a structured conversation around data to network with other data users?
Do you join the conversations?
Are there easy ways to contact the individual ‘data owner’ in your organisation to ask them questions about the data, or to get them to join the conversation?
Are there offline opportunities to have conversations that involve your data?
Whilst the intervening years have seen governments struggle to host open conversations online, and to have the resources to engage with, moderate or manage open discussion threads around datasets, I’m still curious as to how often portals have sought to create more direct channels of communication to data stewards inside the organisation, providing them with feedback on how their data is being used, or offering the opportunity to engage with data users. My working hypothesis is that, if engagement is framed right, many data stewards would be happy to hear from external data re-users and to join conversations about how their data is used.
One notable example of a portal that is successfully hosting significant engagement is data.gouv.fr managed by the French Governments’ Etalab. An ‘Actions’ tab at the top of each dataset page invites users to declare data re-use, or contact the data producer. Further down the dataset page, community discussion space, which appears to have active responses from site moderators, can be found.
With many portals now up to a decade old, the question of how to archive data is becoming more pressing. Data shouldn’t just disappear: but neither should outdated datasets be turning up at the top of search results.
Portals, that started as a snapshot of current organisational data, potentially require new forms of curation over time. This may call for new skills and people contributing to portal management, including archivists.
One paper or provocation I encountered in this rapid review (which I can’t locate again right now) suggested that what we need are ‘question portals’ rather than ‘data portals’ in order to better serve civic engagement. Arguably, this is the role libraries have long-played, with librarians supporting patrons to find the answers they need through information and data. Press departments and FOI teams also play a role in responding to requests for information that may be served by data, but my sense is that these teams are rarely well-integrated with the teams providing portals.
For business and innovation re-use of data in the UK context, Local Enterprise Partnerships and the Growth Hub network could be relevant points of interface to the portal. After all, perhaps data portals should be mainly accessed through people? That is, the portal primarily exists to help the expert user better provide tailored answers to questions from the public, rather than trying to be a self-service public portal for all levels of technical and governmental knowledge.
It is ultimately the people and processes around portals that will determine whether they generate value, and successfully operate as a component of open, accountable government. In any rethinking then of portal futures we need to be sensitive to two possibilities:
1. There’s no place for portals anymore- Portals may have helped us to understand the data landscape of our public bodies, and provided a good data-discovery stop-gap before schema.org and DCAT standards allowed dataset metadata to be harvested from anywhere. However, whilst internal Information Asset Registers might remain useful as management tools, the process of sharing and opening data should shift to departmental responsibilities, supported by capacity building and central support functions - but without the gatekeeping role of a portal.
2. Portals features need rethinking for their future role- Portals remain a relevant tool to take forward organisational data agendas, and promote values of openness and responsible data sharing. They need to be better configured around the people who steward data inside public organisations, and the kinds of processes of public engagement data can support, without trying to ‘host’ all the engagement directly.
In the final post in this series I’ll unpack a bit more some speculative ideas on how portals could be reshaped in future.