By Guest Writer
Open data became a global phenomenon slightly over a decade ago and over the years has been adopted around the world by different stakeholders on the African continent. It refers to data that can be accessed, used, re-used, re-purposed and redistributed by anyone subject only, at most, to the requirement to attribute and share-alike.
The use of open data has been championed for strengthening government transparency, accountability and responsiveness, spurring social and business innovation and fostering inclusivity and empowerment.
Additionally, established open data principles go beyond technical aspects to make its link with socio-economic development outcomes on inclusivity, empowerment, improved governance and citizen engagement. As such, the successful use of open data for development relies on combining the technical aspects of open data such as formats and licensing with the exploration of how the publication of open data contributes to socio-economic development in different contexts.
14 Barriers to Open Data for Development
However, there is little evidence to show that data is being used optimally in informing African countries’ development agenda. The existence of data does not on its own guarantee improved development outcomes. There is still a lot to be done to build data ecosystems that collect and publish data that can adequately inform development objectives.
The Drivers of Data for Development in Africa report found 14 barriers to using open data for development across the continent.
1. Weak legal and policy frameworks
Legal and policy frameworks are key enablers for the use of data in development. Across Africa, many governments, non-state actors and continental institutions are grappling with the legality of data sharing and openness vis a vis privacy and the attendant data protection requirements encoded in statutes. Lack of laws and policies make institutions hesitant to share data for fear of legal ramifications.
This has affected the sustainability of already implemented national open data initiatives, some of which are struggling to collect data from government MDAs due to the lack of a legal and policy framework that provides public servants with cover and which can be used to hold them to accountable.
Multiple laws and or policies are required to fully address the dearth of open data including laws and policies on data protection, access to information, intellectual property, cyber security, open data among others. Most African countries do not have comprehensive laws or policies addressing these issues and in some instances there are conflicting legislations in place.
Where Access to Information (ATI) laws exist, they have provided an anchor for open data initiatives, as in the case of Kenya and Sierra Leone. While these have provided entry points they focus more on information access and have limitations with regards to prescribing formats and licensing for open datasets. This has posed challenges and further entrenched the reliance on champions and political will, both of which are not sustainable approaches in a legal and policy vacuum.
There is also a rapid expansion in the data universe with technologies such as Artificial Intelligence having great potential for unprecedented benefit for society but also great risks for human rights. This has exposed new gaps in legal and policy frameworks for data in Africa affecting their ability to adequately respond to the challenges of the ever expanding technological space. With multiple AI-driven initiatives mushrooming across the continent to address local development challenges, the lack of legal and policy frameworks presents a concern especially with growing anxieties on the exploitation of social media platforms.
2. Lack of binding continental frameworks
The lack of binding continental level guidelines endorsed by key stakeholders, especially the African Union, on the use of open data for development has made it difficult to stimulate adoption at the country level. While the Africa Data Consensus requires that data communities uphold openness, it is not binding and countries can therefore choose to or not to domesticate it.
3. Lack of institutional data policies
Discourses on data policies mostly happen within the context of continental guidelines or country laws and policies. Data for development is however not just a preserve of governments and as such there are many actors involved. While these actors are bound by laws and policies in jurisdictions where they operate, data policies should not only be a feature of data governance in the public sector.
The lack of data policies that allow for data sharing and access at the institutional level presents a barrier to the publication and use of data. The absence of or restrictive data policies reinforce data hoarding tendencies by institutions. Where they are absent, especially within institutions routinely collecting and publishing data, questions can be raised on the security of data collected especially when it has personal identifiable information.
There is a lack of clarity as to what happens to data collected beyond project implementation cycles and how that could possibly impact the safety and security of individuals whose data was collected.
4. Poor data practices
The success of open data for development relies on good quality data. While there is data collected, there is a gap in collecting data that either asks the right questions or provides answers to development related questions. Most data collected does not also meet the fair principles for data management that requires for data to be findable, accessible, interoperable and reusable.
Good quality data needs to be accurate, timely, adequately disaggregated, reusable and comprehensive. Currently most governments and stakeholders only generate reports as PDFs. The lack of raw data and inadequate metadata limits the use and value of published data.
In addition, data published is not adequately disaggregated to the most granular level which decreases its usability since general data/statistics cannot be used to answer specific questions. Gender data for instance is an area with wide gaps that need to be addressed. Consequently, governments traditionally prioritized data on sectors that anchor their economies. Sustainable and inclusive development however requires that governments widen data collection across different sectors.
In addition, data collection and management at national levels are quite fragmented. There are multiple data collection efforts by different government MDAs. This data is however stored on government websites with no inventory of where it can be found. Individual(s) and institution(s) seeking this data find it difficult to locate it thus impeding access.
5. Lack of sustainable funding for D4D efforts
Financing for development data in Africa is mostly available through funding from development partners. A lot of open data initiatives are especially driven by CSOs and researchers who don’t have the budgets to sustainably implement these initiatives due to reliance on short term project based funding.
This lack of funding stifles innovation and makes it difficult to implement any initiatives sustainably. With a lot of funding from development partners, changes in their global and regional priorities/strategies also affect funding for D4D and OD4D and as such support for cross continent projects could stop or be altered without notice.
6. Open for who?
The idea of data for development and more specifically open data is reliant on data being published online. It then goes without saying that for data to be accessed and meaningfully used, users need to first have the connection to access it and then have the right hardware and software to be able to process it or analyze it in different ways.
According to the International Telecommunication Union (ITU), only 28.2 per cent of individuals across the African continent use the internet. Further, access to digital infrastructure and skills to make sense of the data is still very low. Access to open data is therefore not equal and may in fact deepen the ‘digital divide’ as its effective use requires digital infrastructure, hardware and software, financial or educational resources and skills.
7. Imbalances in demand and supply of open data
For data for development to work there needs to be both publication and uptake of the data. With governments and stakeholders increasingly making data open, there still exists a gap in the uptake and use of this data. There are very few existing use cases of initiatives that have
stemmed from published open data. Being able to stimulate the demand side of the open data cycle is therefore a critical part of making D4D work. The lack of awareness of the existence of these datasets is one deterrent to their use. Additionally, the lack of functional feedback loops for users, makes it difficult for them to give feedback on the usefulness of datasets made available/open or to request for specific datasets that they need.
There also have been concerns of stakeholders only making available low value and low political risk datasets that do not hold meaningful potential for informing socio-economic development. There have also been multiple scenarios where institutions attach conflicting licensing to datasets that deters their use. While a dataset may be licensed as open, in some cases the website it is found on isn’t and this makes it difficult for it to be used.
8. Open Government Data Skills gaps
Technical aspects of data analysis require an in-depth understanding the data especially the context in which it exists from a socio-economic development perspective. There exists gaps in the technical skills required to manage, analyze, disseminate, and communicate data to different stakeholders since data literacy is not prioritized in many contexts especially within the public sector.
In addition, most existing models of capacity building favour one-off training over sustained support making it difficult to meaningfully build capacity. It is difficult for stakeholders to master technical skills and concepts over a four or five day workshop.
Finally, while e-learning platforms provide a means for building skills, some skills require more long term training in the form of a university degree, most of which may not be available within some countries and or are not affordable for important stakeholders within and outside of government.
9. Extractive data collection processes by non-state actors
A lot of non-state actors including development partners, private sector and CSOs collect vast amounts of data in the course of their work. However, a lot of this data is not made available to other actors or the citizens who in many instances the providers of the data.
Participants in most cases do not know where the data they provided is stored, how it was used and whether there was any impact from it. Where data is made available, a lot of the data collected is not available beyond the implementation cycle of initiatives.
This also means that different stakeholders are in some cases investing in collecting the same data from the same people at the same time which leads to participants feeling over researched possibly resulting in participant apathy that could negatively affect data quality. This also affects the efficient allocation of resources with stakeholders duplicating each other’s efforts.
10. Few impact stories on data for development
Awareness of existing use cases or impact stories of D4D and OD4D are a huge incentive for further investment by stakeholders. There however have not been many use cases or impact stories for D4D and OD4D within the African context. Where they exist, they have not been amplified enough to reach a broad audience.
In cases where they have, the storytelling has not clearly demonstrated the link between the successes of these initiatives and data use. This can sometimes be attributed to the fact that sustainable socio-economic development is in most cases a result of multiple variables at play, one of which may be the data.
It is therefore difficult to realistically attribute only specific results to the publication and use of data in ways that it demonstrates value for investment.
11. Negative perceptions on data and open data
There is a reluctance by government to embrace open approaches due to safety concerns and fears of perceived ‘misuse’ of the data. Multiple stakeholders also hoard data due to perceptions that it gives them a competitive advantage over others and that it would retain its value if they do.
For researchers, there is scepticism on the value of open data as it sometimes perceived as being low value and that ‘valuable’ resources would only be found behind a paywall. Lack of permanence of most open datasets serves to fan this perception as there is no guarantee that a dataset will be available on the same link years later due to flawed data management procedures affecting its utility in publications due to citation concerns.
12. Lack of contextual innovation
There are and continue to be new technologies supporting data collection, analysis, dissemination and use. Some of the existing technologies may however not be applicable to contexts across the African continent. In addition these available data solutions may not also provide room for adjustments therefore making them difficult to use. Local innovation would serve to bridge this gap by developing context specific tech solutions that take into consideration the legal, structural and cultural realities of different contexts.
13. Lack of coordination mechanisms
Multiple stakeholders working on the African continent implement initiatives that are either similar or that complement each other’s efforts. The siloed working structures make it difficult to consolidate advocacy efforts, prioritize areas of investments and has resulted in the duplication of initiatives, data collection and publication efforts with multiple dashboards featuring similar data.
14. Reinforcement of existing structural inequalities
Structural inequalities permeate different aspects of society such as imbalances in economic power that put certain groups/individuals at a disadvantage.
There have been concerns by stakeholders, both in the private sector and civil society, that opening up their data exposes them and or their data to exploitation by bigger actors with more resources within the same space as them. This has therefore meant very little recognition and value for their work providing no incentive for them to keep publishing their data.
Additionally, approaches used to collect open data such as crowdsourcing significantly use the labour and expertise of volunteers much of which is either not recognised in the larger dataset and does not translate into any value for them.
