'Peer learning network' banner

Stewards of open data: what impact have they had?

Thu Dec 16, 2021
$download_content = get_field('download_content');

We share insights from our ‘Stewards of open data: What impact have they had?’ roundtable, as part of our second-peer learning network in partnership with Microsoft

At ODI Summit 2021, we held a roundtable to explore organisations that steward open data to unlock value and the impact that these organisations have had, as part of our second peer-learning network with Microsoft. In this blogpost, we share recordings from the roundtable alongside insights from other members of our peer-learning network. 

Unlocking value from open data

Data exists on a spectrum, from closed to shared to open. By making data as open as possible – while protecting people from harmful impacts – we can unlock more value from that data by creating more opportunities for it to be used in innovative ways.

The Data Spectrum illustrates types of data access, from closed to shared to open. Moving from closed to shared to open, the types of access are: internal access; named access; group-based access; public access; and anyone.
The Data Spectrum illustrates how data ranges from closed to shared to open

There are a number of organisations that steward open data in order to create value, or to support others to do so. As part of the ODI’s second peer-learning network, in partnership with Microsoft, we hosted a roundtable to explore the work that some of these organisations have been delivering, and to discuss what impact they have had. The speakers on this roundtable were representatives of MetaBrainz Foundation, Open Apparel Registry and OpenStreetMap Foundation – three organisations who are part of the peer-learning network.

In this blogpost, we share the findings of the session through recordings of the roundtable. Below each clip, we’ve also included written responses from other peer-learning network participants who did not participate in the roundtable – the Global Biodiversity Information Facility (GBIF), Open Contracting Partnership’s (OCP) and WikiRate.

Stewards of open data: what impact have they had?

Introduction

Question: Why did your organisation opt to become an open data steward?

A common purpose among our peer-learning network members is the idea that openness improves accessibility, which can be beneficial for a number of reasons. Here are some examples from the group:

GBIF: GBIF focuses on making biodiversity data openly available. It was created after a 1999 report concluded that ‘An international mechanism is needed to make biodiversity data and information accessible worldwide’, arguing that this mechanism could produce many economic and social benefits and enable sustainable development by providing sound scientific evidence.

OCP: OCP’s vision, objectives, activity areas and general structure were created through a collaborative process with around 200 diverse stakeholders from the public, private and civil society sectors. Many of them had already recognised the potential of open contracting and were working on contracting-related issues within their organisations or sector, but in topical or geographic silos. To address the lack of coordination, stakeholders at the meeting enthusiastically embraced the idea of establishing the OCP as a platform for collective action and learning.

WikiRate: WikiRate was founded as an open data project because the main issue in the corporate sustainability data space is that ‘good’ data (clean, standardised, available at scale) is stuck behind paywalls. Everyone should have access to ‘good’ data on company impacts so that everyone can weigh in on the roles companies can and should play in society. WikiRate aims to democratise the access to and use of company data so that everyone can make more informed decisions.

Question: What does impact look like for you in your work? How do you measure it? How do you maximise it?

Impact for open organisations can vary depending on a number of factors, such as organisation size, sector and audience. For example:

GBIF: GBIF’s ultimate vision is ‘A world in which the best possible biodiversity data underpins research, policy and decisions’. At the highest level, GBIF looks for impact on policy and decisions to address biodiversity loss and sustainable development, which build on the best available scientific knowledge. The open data available through GBIF is primarily targeted at scientific researchers who deal with questions relating to species and their location. Most direct users are the scientists who reuse the data GBIF assembles from a wide network, but true impact is realised through a culmination of the work produced by these scientists, and others who might use the data.

OCP: OCP’s vision is that fair and effective public contracting provides everyone, everywhere with the public goods, works and services that they need. It delivers on this vision by working with governments to open up and transform contracting, to make it better and fairer. For concrete examples of impact created by OCP, you can read its impact stories (filter for ‘Impact Story’). You can also read more about the impact of open contracting.

 

Measuring and maximising the impact of open data can be difficult. Here are some examples of how the organisations in the peer-learning network approach measuring and maximising impact:

GBIF: GBIF has found that it is easier to measure and track the more direct impact of the use of biodiversity data, mobilised through GBIF in research, than it is on policy and decisions regarding conservation and sustainable development, where impact is often indirect. Over the last few years, GBIF has developed processes to enable data published through GBIF to be cited. To date, it’s tracked 6,000 peer-reviewed papers that use this data, and are seeing new citations in approximately 3 to 4 new papers per day.Tracking impact in policy and decisions is more challenging, but GBIF are aware of its contribution to intergovernmental processes that lead to global policies on biodiversity. One example is through the United Nations Convention on Biological Diversity, which uses data from GBIF to contribute to indicators used to measure progress on the globally agreed targets.

OCP: OCP can see impact within a single government agency or across multiple agencies, as long as the change occurs across a large number of projects. Because these are significant changes, impacts are usually seen two to five years (or more) after an intervention begins. To count as impact, OCP must be able to verify the quantitative results through rigorous evaluative measurement, and the impact needs to be attributable to some combination of our four cornerstones of open contracting: user-centered design of reforms; open contracting data; cross-sectoral engagement and feedback; and learning, sharing, and iteration. To maximise impact, OCP embeds monitoring, evaluation and learning (MEL) in its work: see its learning plan. It also conducts an annual survey of its stakeholders to ensure it stays on track. See its 2019-2023 strategy and 2021 strategy refresh for a high-level view.

Question: What challenges have you found in making open data impactful?

Open data stewards often face a number of common challenges. Some of the examples that our peer-learning network members have shared include:

  • Ensuring that data publication occurs at a scale to allow the infrastructure to provide access to data to meet users’ needs. This involves a range of interlinked challenges on the technical and cultural levels, such as the adoption of sufficiently simple data models and tools for data publishing and access that can still be adapted to the needs of multiple users.
  • The perception of what should be considered ‘quality’ data. What is deemed as data of poor quality for one purpose may be valued greatly by another consumer.
  • Governments sometimes treat transparency as an end in itself. A big challenge is to ensure that data is being published with attention to serving real, expressed needs.
  • Governments sometimes publish data but don’t use it themselves. This puts the publication at greater risk of discontinuation.
  • For a publication to get off the ground, initial challenges include finding the political will (‘air cover’), bringing champions on board, securing resources, etc.
  • Many stakeholders have low capacity to use data, especially contracting data. The challenge is to provide the learning resources and data tools to ease use, and to nurture a self-sustaining user community to help each other.
  • Complexity of the data has been one of the key challenges to making the data impactful. There is a simplicity that is assumed which just isn’t there in reality.

Question: Where can other aspiring open organisations go to find out more information about developing their business models?

A number of our peer-learning network members have used resources made available by the ODI. Here is a list of resources that may be useful to organisations which are interested in exploring open data stewardship models:

In addition, the ODI and Microsoft will be hosting a roundtable discussion in January 2022, with the peer-learning network, to discuss how organisations develop a sustainable business model without compromising commitments to open data. Learnings from this roundtable session will be shared publicly.

Question: How do we get the private sector to share data? And how do you get them to tell you what they are doing with the data you publish?

Here are some ways in which the peer-learning network members engage with the private sector:

GBIF: GBIF actively funds projects to engage and promote sharing from private companies – for example the OpenPSD project.

OCP: OCP believes that there are many ways to get the private sector to share what they are doing with public contracting data. Some methods are passive, like media monitoring and web/API traffic analysis. Other methods are active, like developing relationships with private actors, offering to communicate their achievements, supporting the types of private activity that support your mission, or offering your communications platform if their message is in line with your mission.

Question: How do you define open within your organisation? Are you using data that sits elsewhere on the Data Spectrum? How do we move away from open washing?

The term ‘open’ can be interpreted many differently by different people, so it is important that organisations that steward open data are clear about what they mean when they talk about openness and open data.

GBIF: GBIF uses Creative Commons open licences to outline a set of clear rules around reuse of the data it publishes, and it uses the FAIR principles of Findability, Accessibility, Interoperability and Reusability to guide its work.

OCP: OCP uses a number of different sources to talk about all things open, such as:

OCP only uses public data, which sits in the open and shared parts of the Data Spectrum.

In order to avoid ‘open washing’ (claiming something is open when it doesn’t meet a widely accepted criteria for openness), OCP has a specific criteria about whom it engages with. To access OCP support, partners need to demonstrate that they are adopting open contracting to achieve a meaningful outcome, like improved access to medicine – rather than adopting it to give a veneer of ‘openness’ to their work.

That said, all OCP resources are openly licensed, so it remains possible for actors to open-wash without its direct support.

Question: What would your advice be to other organisations considering becoming open data stewards?

GBIF:

  • Pay attention to data quality at every step, and quantify it to set clear expectations for consumers
  • Consider the comfort level of those sharing data
    • a tiered approach may be applicable, where people can progressively open up the richness level of their data
    • Find mechanisms that bring real value to the publishers
      • For example, GBIF promotes and assists getting a peer-reviewed ‘data paper’ in a journal
  • Pay attention to clear attribution/citation and licensing, and consider mechanisms to track use
  • Developing community norms openly – involve the community in all aspects around data publishing, use and technology

OCP:

  • Focus on serving the real needs experienced and expressed by your stakeholders.
    • Include user research and usability testing in all your work. (Do not take a ‘convenience sample’ when selecting subjects!)
    • You can get lost in the technical aspects of stewarding data, and in the never-ending task of improving quality and completeness. Always try to bring the abstract value back to the concrete goal.
    • Be sensitive to the technical capabilities of your stakeholders. If they are at the level of working with spreadsheets, you might not want to publish JSON-LD.
    • Using the data yourself can also give you insight into your stakeholders’ challenges – but don’t rely on it too heavily!
  • Stewarding data is not only a technical exercise. You also need to invest in learning materials and/or training events to increase people’s capacity for using that data effectively, and in community building to motivate use and encourage connections between stakeholders, to support and build on each others’ work. See OCP’s community building strategy.
  • If you do not ‘own’ the data but rely on others to publish it, you will also need to have a plan for advocacy and communications, to convince data holders to publish openly and to maintain momentum for continued publication.
  • Some of this advice is the same that we give to publishers of open contracting data. See, for example, OCP’s Quickstart Guide and Open Contracting Playbook (you can skip the content that is specific to procurement).

WikiRate: 

It is important to think of your (potential) data users in ‘tiers’.  Are they direct users of your data or are they indirect users? Say the data you have are GPS coordinates. A first-tier user of that data would be an organisation that has developed a navigation app and the second-tier users are those using that navigation app. Their knowledge and understanding of the data (and possibly tech tools) that you have to offer differs greatly, and you need to make sure you understand both use cases so that you can integrate their needs into your system and the data that you provide access to.

This is the first of a number of resources that we are producing about the experiences of the open data stewards which are participating in our peer- learning network. If you are considering adopting an open data stewardship role and you would like to keep informed about the outputs from this project, which will cover a range of topics such as organisational governance, sustainability and data access, please get in contact.