Open Data for Qatar: Now What?
The Qatari Ministry of Information and Communication Technologies announced a finalised Open Data policy in December 2014. It is encouraging to observe that, despite its various transparency caveats, Qatar is quite openly discussing an Open Data policy.
A year ago, I wrote a quick overview of Qatar’s draft Open Data Policy paper. Alongside, I also responded to the public consultation on how this policy should evolve. The Qatari Ministry of Information and Communication Technologies (generally referred to as ictQATAR) announced the finalised policy in December 2014, through a dedicated one-day workshop.
Naturally, I was intrigued to see how the text has evolved. Not only because I contributed to the public consultation. But also because the Open Data Barometer rankings were recently published: Qatar features amongst the ‘one-sided Open Data initiatives’ there, alongside Saudi Arabia and Bahrain:
These countries each have some form of open data initiative, ranging from departmental web pages that display open data, to full open data portals. However, government action to publish selected datasets is not matched by civil society capacity and freedom to engage with the data, nor by private sector involvement in the open data process. As a result, these initiatives appear to be very supply-side driven, without engagement with a broad community of users. Without wider political freedoms, the potential of open data to bring about political and social change in these contexts will be limited.
The image on the right is even more telling. It indicates that government readiness, albeit non-zero, remains pretty limited. The same is true for entrepreneurs, but these are the best ranks the country actually gets. Indeed, evidence for neither citizen engagement with Open Data was identified nor was the data used to bring about some flavour of change in society and economy.
What also caught my attention is the evolution of Qatar’s performance in terms of Open Data. The Barometer’s first edition assessed 2013, so it is easy to compare 2014 results. Qatar’s evolution is surprising: it lost four ranks in 2014 compared to 2013. Also, although government and entrepreneurship readiness levels have slightly increased, all the other indicators have sunk.
The Barometer formulated the need to introduce “wider political freedoms”. If present, these would enable Open Data to become a political and social game-changer. This need is even more striking as the policy draft did not make any mention of it. Yet, such an urge is corroborated in a totally independent fashion by a report released on 26 April 2014 by private firm Tahseen Consulting. It provided several insights that can enable a greater development for Open Data initiatives. One of these recommendations was to increase citizen participation:
There is little evidence which suggests that civil society or information technology professional groups are being actively engaged by governments in forming open government and data strategies, identifying data requirements, or in increasing citizen use of open data. Open data initiatives are generally designed to broadcast data rather than create a genuine dialogue about what data might be required by the community. There is a significant role that civil society can play in defining the types of data citizens might find useful and in analyzing publicly available data.
The blog post announcing the finalised Open Data policy featured on ictQatar’s National Strategy page. Yet, the country does not seem to showcase a clearly defined governmental support to open data driven innovation. The ictQatar has identified four different digital strategy axes but these seem to bolster services and infrastracture.
The ictQatar is aware about what challenges various Open Data initiatives around the world have faced. Along with the above considerations, I went into closely examining and comparing the draft and the final policy papers.
What has changed
The proposed article 4.2 (iv.) was removed in the final policy. It stated:
The proposed national Open Data portal shall operate as an index to serve as gateway to public for dataset discovery and search, and shall redirect to respective Government Agencies’ data source or webpage for download. The portal shall also aggregate all public feedback and requests, and the government agencies’ responses to the same
It seems to say that each agency will need to maintain its own Open Data portal rather than relying on a national one. This is at least what I gather since no clear mention is made of a centralised national-level Open Data portal.
Good to see that art. 4.4 (ii.) asserts that,
Access, use, adaptation and distribution of the Open Data should be permitted at no fee to the public.
And even better to see that art. 4.4 (iii.) clearly asserts:
Individuals must not be required to register or identify themselves in order to access, use, adapt or distribute the Open Data.
As a comparison, to download Saudi Arabia’s Open Data Guide, you have to provide personal details and login. The free (as in free of charge) and open (as in no login required) access to data are reasserted in art.4.7 (v.) and (vi.), while they were absent from the draft policy.
With regards to requests for data, the final policy has considerably improved one provision: namely, the delays one may face after requesting data to be opened up. Thus, art. 4.5 (iii.) reads:
iii. Agencies shall respond to all feedbacks, suggestions and Information Requests received on a timely basis and within a period not exceeding 15 (fifteen) days of receipt of the request. This does not include the time needed to prepare the information or data requested. However, their response shall include whether they shall be able to provide the information and the time required to provide such information. Agencies shall use the online mode available on their Open Data webpages in publishing their responses. [this is has been appended to the final policy statement]
But then, the good ol’ idea that we can also make money by selling data comes back… Indeed, art. 4.4 (ii.) stated that nobody should pay to access the data. Yet, art. 4.5 (vi.) enforces:
Agencies may charge a reasonable fee from the users to release data (not applicable on voluntary release per 4.6) against requests noting that such data may not be classified as Open Data or may require additional efforts not already planned or budgeted. Agencies shall establish and publish a datarequest fee framework that shall apply in case of such requests.
This provision was absent from the draft policy. I am not aware of a situation where this has materialised elsewhere: if you have such an example in mind, holer at me.
What is missing
The policy seems to take Sunlight Foundation’s Ten principles for opening up government information as a reference scheme. But a few of the key components are missing from Qatar’s Open Data policy. The latter is in general written in fluffy language and some of the provisions are too vaguely worded.
I found no mention of specific datasets to be opened up. The policy paper highlights the key datasets as identified by G8 Open Data Chapter (Appendix 4). The public consultation on the draft policy included a dedicated question (Q8), to which I formulated the following response:
Pollution levels, energy consumption, governmental spending, contracting and procurement (call for tender, future tenders, etc.), local budget, national budget (planned and spent), education and health related datasets (performance, prescription of public institutions), transport and infrastructures, company registrars, etc. These are both relevant for the population and allow for an improvement of the public sector transparency and government accountability.
From what I gather, each Agency decides on its own what to open and what to withhold. Moreover, the policy poses that “Agencies shall share non-personal data (subject to privacy, confidentiality, security, trade secret, contractual, or other valid restrictions) with the general public”. The lack of clear identification of datasets along with contractual data and “other valid restrictions” leave room for a great deal of public sector data to never be published.
Another shortcoming is the lack of details about licensing (mentioned as a key principle by Sunlight). I have addressed this particular point both in my response to the public consultation and when writing up the companion analysis on this website. The final policy, however, remains unchanged on licensing. Yes, the checklist in Appendix 6 provides general guidelines aiming to inform whether a dataset is releasable as Open Data, and copyright features therein. But clearing a dataset from third-party ownership and/or from having granted an exclusive license to another party is far from enough when it boils down to defining a clear licensing framework for the forthcoming Open Data portal.
Another point about technicalities is the data publication itself. Regardless of the government agency, some datasets will need regular updates. These will thus require repeated publication rounds (think of spending for instance). How will these be published? Will the agencies release an Excel spreadsheet for every update? Speaking of spreadsheets: the point about formats is totally unchanged, despite my explicit question in the response to the public consultation. So far, the policy makes no mention at all about formats. Finally, there is strictly no mention of high-quality dataset organisation (URIs, linked data, etc.).
I already highlighted the urging need to involve citizens and businesses (i.e., the ‘demand side’, to use a policy-maker parlance) in this process. Reading the policy does not, however, inform me on whether such an effort will be made. Indeed, the policy does not spell out any recommendation or commitment in this direction. Both the timeline (Appendix 8) and the Policy Compliance Checklist (Appendix 9) strictly refer to data release, but no mention is made of citizen engagement.
Although government shows some Open Data readiness, its level is far from satisfactory. Yet, I fail to identify evidence that training on Open Data is available and/or will be proposed anytime soon. The Policy Compliance Checklist (Appendix 9) makes no mention of training on Open Data for these agencies either. Indeed, some general data science courses exist, but they are insufficient. This shortcoming was also highlighted by Tahseen Consulting’s aforementioned report:
Relative to OECD countries, most GCC governments focus on national data portals rather than regional and city initiatives, have not enacted right to information laws, fail to engage civil society and academia in efforts, and lack education and training courses for developing more effective open data programs.
At last, the final policy timeline mentions that a “Senior Official with sufficient authority to assume the Open Data responsibilities” will be designated within three months from announcement of the finalised policy. This person will be in charge of government agencies efforts towards Open Data. Yet, no such nomination seems to have happened nearly five months after the announcement of the final policy.
To sum it up: Qatar has produced a vaguely worded paper which lacks major tangible commitments and has no mention of the most pressing yet basic requirements for a successful Open Data initiative. Each Agency is left to decide what data to release. Since the country has no right to information legislation, it is difficult to predict what will be coming out from this initiative. You have understood it so far: I am wondering what the strategy behind this policy is.
Openness is more than just infrastructure and policy papers.Posted on: April 16, 2015, by : rayna