12/1/13

Open Development & Open Data - Challenges and Opportunities

Global Development Professionals Network in London ICT4D Meetup:

'Open Data in Development' and 'Open Development' has sort of come of age. Global financial institutions, recipients of aid like Kenya, donors like DFID and the Bill and Melinda Gates Foundation see potential in the arrival of open data to the development world, as posts by Chris Gingerich and Saara Romu indicate

The World Bank has recently announced a formal partnership with the Open Knowledge Foundation and the Open Data Institute in London to work together over the next three years. In addition to this, another World Bank initiative called the Open Development Technology Alliance (ODTA) already has 7 draft reports out on the subject.

Organisations such as Publish What You Fund and Ushahidi have been doing it all since 2008. The Open Knowledge Foundation (2004) also has supported these efforts and many others, acting as the broad umbrella that brings open data people together, coordinating working groups on 'Open Development' as well as one on Open Sustainability with blogs, and resources like this simple video from Open For Change. It describes how opening up different forms of content, using licenses and standards can generate new forms of collaboration to achieve different goals in development work. Open for Change was also behind the Open Development Camp 2013 which recently took place in Amsterdam.

Harnessing open data to be used for development work may bring significant benefits. Overseas aid and humanitarian work could be made effective by using data to challenge many of the assumptions that are often wrongly made about developing countries and disaster situations. Data can be opened up for different groups to use in different ways to help tackle corruption and may even encourage policy making to be based on harder evidence. High expectations for those involved in the delivery of aid to be transparent and accountable are most likely to increase, not diminish. For example, this might have implications for a body such as the Global Fund that is being accused in the British press of having watered down its own reporting and auditing proceedures.

Even broader questions though were posed at this camp in November, to be considered. How can open data, innovative technology and new ways of working help us create a better world? How relevant is open data when you live on less than $1 a day? How can we avoid yet another divide between the haves and the have nots in this shift? What are the security risks?

NGO’s and foundations working in development are also not the only ones getting involved. The UN has also been working on opening up its practices and its data. The UNDP for example has recently launched its new open data platform and work is continuing on other projects such as the Common Humanitarian Exchange Language which is coordinated by the humanitarian agencies.

Recently at a meetup on the 26th of November, a group of 58 people at the Mozilla offices in London met to dig into Open Data for Development. If you missed it and would have liked to have come, there will be more events in the future and the OKFN will hold a big conference in Berlin in July. The event was organised by the London ICT4D meetup group and by Eliza Anyangwe who runs the Guardian Global Development Professionals Network. This is a summary of the discussion, along with some observations. Please do add feedback to this post here in the comments section below, and any comments that you wish to add are very welcome.

Eliza Anyangwe brought along an engaging panel, consisting of Ulrich Atz (Open Data Institute), Heather Leson (OKFNlabs) and Claire Provost (Guardian Global Development) who were asked to address two questions:

- What are the hopes and aspirations for open data / open development
- What are the challenges and obstacles ahead?


                                                                                                Photo by Daniel Fowler

A good contingent of ICT4D researchers also attended along with doctoral researcher Tony Roberts, spiritual sherpa and founder of the ICT4D London meetup group with a squad of researchers from Royal Holloway. The academic community were well represented with Duncan Edwards from IDS (who has written on this topic here), Vanessa Thomas from the HighWire Doctoral Training Centre in Lancaster, Gregory Grisha Asmolov from the LSE and Geraud De Ville from the Open University. It was also great to have representatives from many organisations and NGO’s including Aptivate, One World, the IRC, Africa Gathering, CIFOR, Publish What you Fund, and of course the OKFN.

Eliza Anyangwe has written a recent piece for the Guardian on this topic, and kicked off the proceedings with a provocation looking back at Florence Nightingale’s use of data visualisation techniques in her own campaign for changes in health treatment in the British army. Not only did her complex and innovative graphs present rich data, they were also constructed to prove a point. Her analysis and representation of the causes of death and disease bordered on advertising for her cause. With data visualisation now having become so ubiquitous in journalism and in the media, are we getting to a point where we are simply losing the ability to distinguish between information and misinformation? If open development work proceeds in a similar way, what happens if the data quality is not addressed adequately? What are the things getting lost in the glare of data visualisation?

Heather Leson from the OKFN stressed the importance of having the right kinds of feedback loops both in terms of feelings and methods. Feelings count because they indicate how to build trust in the sources, or whether people understand the data. People often don't trust the data unless they are involved in the data collection. Until we talk more about how we involve citizens, then we won't know how to get the right feedback, the remixing and re-use of data that helps to get a more accurate 'open data picture'.

In terms of methods, she pointed out that standards for interoperability are important, referring to a clean Dataset Guidelines list. She pointed out how important it was in an election project with Ushaidi to have communities involved in cleaning data in real time. Examples were also given of data cleaning projects using machines to do this.

Finally, Heather focused on the need to teach and build data literacy. The School of Data has held experimental training workshops for early adopters already. We now need to think more about the question of how to cross the gap between these initial efforts and reaching the next 1000 datamakers in civil society and government.

Ulrich Atz, the head of statistics at the ODI provided some further concise reflections. He pointed out that it is not just about making data ‘beautiful’ but that ‘openness’ involved addressing specific technological, financial and legal hurdles. He made the point that bad data also stifles innovation, leading to problems for sustainability of the whole enterprise. The lack of standards creates inefficiencies that create barriers to data re-use. He pointed out that we need to think about more than just transparency, and consider the role of data within the whole ecosystem. Before it becomes fit for purpose, it is important to test the quality, reliability and validity of data.

He focused on the power of standards and on the ODI’s work on creating certificates for open data. This is a kind of badge that links to a description of your open data. The description explores things like how often it's updated, what format it's in, who and where it came from. For more information see the ODI’s work on this here.

He also made a call to action for us to collect success stories and to remind people to keep holding governments to account for the pledges that they have made to make their data open. He pointed out that a lot of attention had focused on the potential social value of open data, but relatively little had focused on the environmental or economic benefits and possibilities.

Claire Provost who works for the Guardian on global development issues made several points from the perspective of someone at the sharp end of using open data for journalism in the public domain. Timely data can come at the cost of better quality data, and there is often an easy conflation between data journalism and visualisation. Learning to apply the same journalistic scrutiny to datasets as other sources is crucial, and tools that assist in making the process of evaluation of datasets dynamic (such as the aid transparency tracker) are very useful. This tool gives you the ability to understand the plans from various organisations on publishing data, the data fields across all their plans, and gives an analysis of their commitments to publish aid information. Another potentially useful tool mentioned in discussion was also the P2P search engine Yacy. As Clair put it, for the investigator it helps first to know exactly what you don't know.

From her perspective as a journalist it seemed that high stakes questions about the right to information still often are where to start digging. There are many situations in which a well placed request for information, the process of demanding the right to certain data for a story is often the most effective way of fact checking and getting the story right. At least you'd assume the process itself reveals a lot about who is and who is not being given access to data.

The group discussion that followed touched on many points, with journalists, activists, and data scientists approaching the same questions from different perspectives. But here are some of the major points: .

1. Negatives:
We do need to look at negatives as well as positives, and look at situations where open data is not working well. Rufus Pollock at OKFNlabs has recently posted on the topic of Bad Data here.

2. Ethics:
We need to consider ethics as well and realise that in different fields ‘open data’ practitioners can be led into high stakes conflicts, such as on land title issues. An interesting perspective on ethics, informed participation and informed risk-taking is also been posted on the OKFNlabs blog here.

3. Experiment:
We should encourage people to try and experiment and see what happens. Mark Brough pointed to work he had just done on tracking aid to the Philippines here and to the potential for quick, agile and effective action by data wranglers of all kinds in helping with these post-disaster coordination projects in a short time span.

The Center for Global Development goes into more detail about how open data initiatives such as this and FAITH, the Philippine government's own Foreign Aid Transparency Hub launched 10 days after the crisis have played important roles already in helping to improve coordination and effectiveness.  

4. Local Knowledge is important:
We should bring data to local partners to critique and use in different ways, to allow them to make sense of it from their perspective. This point seemed to be about looking for more participatory and inclusive approaches to development using open data. It is a gap which early on was picked up by the UNDP. In this area the UNDP has pointed out the importance of creating channels for people’s voice within institutions in order to make them more responsive to the poor.

5. Think about goals:
Do we need to return to thinking about the goals of development rather than always talking about the means and the tools? If we are thinking about opening up development for the future, then for what and for who? Should we still be calling it development at all? Is using a framework and language of development helpful in terms of dealing with the problems of climate change, migration and systemic risk today? Or should we be attempting to subvert or re-cast those terms?

6. Don’t be a bubble:
Open data for development claims to be a potential force for good, but if it remains both mainstream and expert driven it may not evolve to close the widening gap between the highest levels and people who are the most disempowered and decision making at a local level. In terms of dialogue, we are still a little bubble. People find exciting ways to circumvent that insularity continuously, but unfortunately it is part of the DNA of open development that it lives in the hands of technical experts. Over time concerns such as building linguistic localisation, creating clear documentation, using simple language and searching for criticism and making openness a cultural watchword for how we operate, may need to be addressed more fully. The point was made that the open data movement is still relatively new and has only just emerged to find itself in the mainstream but from another perspective this problem is not at all new. A dependence on technical expertise, closed language and forms of knowledge has always featured heavily within technical communities and development communities alike.

Quick wins for open development in summary: 
- use local partners: they know their contexts best
- integrate with media sources to build on existing more popular media
- train each other to analyse data
- fill gaps in terms of standards and semantics
- call out institutions that do not make data interoperable
- remind governments of their commitments
- publish metadata and the methodology of what gets published
- share examples of what open data can change
- share failure and be open about what is not working
- be realistic about what the small community can solve. This movement is still fairly new
- allow people to tinker with data 

Further comments

Jay Naidoo, Chair of the Board of Directors and Chair of the Partnership Council of the Global Alliance for Improved Nutrition (GAIN) made a clear call in his keynote at OKCON in 2013 for open development to bring politics to the table, for a 'revolution of morality'. He urged us to think about the ethics, to think about how we mobilise citizens as whistleblowers and journalists, to think about ways to also bring about the progressive realisation of human rights through how we are using technology.



After the panel I decided to reflect on this call to action and to the idea of bringing politics in. How would you create tools to build a more caring and humane society that focuses more on the vulnerable? Where should politics be more involved in open development?

Political discussions about goals, mobilisation and even ethics are often separated from the discussions about standards, practices and 'lessons learned'. Since the High Level Panel's report on the Post-2015 agenda coined the phrase, the call for a 'data revolution' has been moderately anodyne. Open data work seems more focused on tools, distilling expertise in the manipulation, statistical calculation and collection of data than on the kinds of stories we are seeking to tell and to whom.

But I’d like to look at three reasons why it seems to me people should do more thinking about linking these seperate discussions together, at least in open development. I'd like to draw out some of these questions for people to wrangle with in the open community.

The first reason is that open data is not really seperate from existing reforms to public service delivery and governance. It could just be seen as the logical extension of other political reforms to decentralise government and decision making. Both the left and the right have their different ideas about this and tend to see different types of opportunities in decentralisation, and so also in 'openness'. For the left it is an opportunity to open channels for giving voice and bringing processes closer to the people, while for the right it appeals because it opens up accountability measures and short-cuts, providing opportunities for improved efficiency and cost savings. It can sit politically within a 'third way' approach. But either way, it will need to fit in clearly with the designs of politics in any given context.

The second reason is that it is implicitly a challenge to the status quo from the 'outside'. A dynamic of innovation seems to be building on these new insider / outsider relationships. Decision making is concentrated in the hands of those that hold power, experience and wisdom. The proof of which is in their embodiment of assets such as knowledge and networks, implicitly shaping discourse. The mainstream if you like is geared powerfully towards sustaining and protecting all that it has created, and less towards change and the new. In contrast the relatively fragile, emergent and innovative solutions adopted by outsiders and different actors are partly designed to disrupt this. There are always ‘early adopters’ of change, and a necessary and sufficient point of convergence for change. To be clear if the 'data revolution' is about a change process and innovation, handing power over in order to bring about change, then within this process there will always be different pathways and circles driving it.

But the point here was to be critical. Providing critiques of open data initiatives is a necessary and essential part of stirring the pot in a process of evaluation and reflection. Bringing attention to the fact that there are different understandings and versions of 'open data' out there is healthy and opens pathways to improve what people do. This debate between Rob Kitchin and David Eaves is one example.  

It is worth mentioning the wonderful opportunity to learn from others, particularly by reflecting on what stands at the heart of being 'open’ for different communities and where that leads to. One definition proposed by Mitchell Baker, Executive Chair of the Mozilla Foundation, to me highlights the potential for innovation and problem solving. 


She suggests that by building things in a way that they are inherently interoperable, we have the opportunity to try new things and so we can do more. By documenting and making things knowable and transparent, we can see them and understand them, so we can know more. When we do more and know more, she suggests there are many other things we can do better. The hope might be for development that this will foster and build accountability and trustworthiness, and individual choice and empowerment. Openness is a mission that the open movement is passionate about and is very good at. The good news is that there is real expertise to build on and learn from everywhere within this movement. The work of Open Matt on openness at Mozilla for example, is a shining demonstration of this.

Thirdly, this seems to me to be political because the future of development work itself is at something of a turning point. Development work in the future is likely to be different, possibly requiring different kinds of political agreements between new and old donors, emerging middle income and lower income countries. At the same time that we are discussing new priorities for development, some donor countries such as Australia and Canada are locking down their bilateral aid agencies, while others are heralding the end of the golden age of NGO's, or limiting their funding for example in Kenya. Many governments seem to be keen to 'leave room' for wider and leaner action by the private sector or initiatives driven by corporate social responsibility.

The point here is to highlight the danger of simply debating the tools and the means, without getting clear about where we are going and how. Planning for the future is also a process of working backwards from a vision. But there are some quite different versions of that vision out there at this current time.

Nancy Birdsall for one, sees this as a simple and straightforward leap towards the end of aid and development altogether by 2030. However, her ideas detailed in her blog post '2030 ODA No More' stand in contrast to those placing greater importance on tackling inequality (see the Guardian's debate on this or the IPS). Other perspectives such as that of Amartya Sen from the capability perspective, are aiming for a more human centered and rights based model for development. To be clear, there are likely to be at least as many different understandings of ‘open development’ as there are of ‘openness’ and ‘development’ alone.

Finally, calling something ‘open’ doesn’t just make it so. Development is a discourse that has a history spanning over many years. It is a world in which global institutions largely have been able to impose policy and change from above. The servicing of debt to global lenders with structural adjustment policies throughout the 90’s for example was imposed on countries in the process of fostering 'modern institutions, democracy and economic growth' in the process of ‘development’. It also led to the imposition of many destructive policies. 

The simple fact of encouraging participation of different voices and ideas now does not simply lead to all voices counting equally. Participation and openness are not the same thing either. As Tiago Peixoto points out in his piece on experiments with participatory budgeting, participation is always unequal. We should consider inclusiveness in our approach to ‘openness’, promoting the participation of previously marginalized sectors of society. The amplification of networked and distributed decision making alone does not undo the fact that power is overly concentrated in the hands of a few. It may simply mean that those who control such networks have a mechanism through which to speak for the multitude across a new architecture of interconnected nodes with greater complexity and power. 

Or as Ruth Carlitz suggests, is participation really being encouraged or is it simply a framework for ‘being participated with’. Surely our standards for ‘openness’ have to take into account in the end whether the dynamic suggests real participation or not? And to what extent might this dynamic be reflected within global transparency initiatives? For example while governments in many countries have been encouraged to become more 'open' about their budgets and their aid giving, the global lenders at the IMF score comparatively poorly on ‘openness’ and are not yet IATI signatories.

Without bringing attention to some of the political issues, efforts to widen decentralised governance and accountability through ‘openness’ will not take into account the political questions highlighted in the points above. That is to say, how such a process gains traction from both the left and the right in any context within political institutions, how it implicitly articulates itself as a challenge to the status quo, and finally how it conceives of goals and leads towards a vision for the future in order to confront problems such as widening inequality, financial risk and unmitigated climate change. ‘Participation’ in short is not enough, it simply renders this 'openness' in the weakest sense.

Tiago Peixoto has also pointed out that transparency projects are also often the “low-hanging fruit” of open governance, and are tempting for governments to focus on. His work has questioned the simplicity with which notions of participation and participatory experiments have been thought to lead directly to more effective collective action. Without a definition of ‘openness’ in development that also considers political barriers explicitly, that also encompasses barriers to collective action, there is a real sense that this ‘openness’ serves partly to isolate and close off aspects of this area of decision making from view. What remains outside of our own definitions of ‘open development’ should concern us, or we may end up refashioning development for its own purposes more than for anyone else's.