Open Data – 2014 Digital 21 Strategy – Public Consultation

Dear Gregory So and the OGCIO,

I wished to add some perspective to the role that open data can play in the Digital 21 Strategy. The passage below from the public consultation provides a very simplistic portrayal of how every government should proceed when it comes to open data. This statement does not go far enough. The success of any open data strategy is dependant upon whether it gets embedded in the culture of how all Government Departments function. It is not sufficient to say that data will be released in machine-readable formats. Instead, there needs to be a greater emphasis on the idea that much of the information that is produced by the Government should be publicly released immediately (produced to be made public). The position that Data.One has taken is exploratory at best. They seek to find ‘high-value’ datasets and help release them. This endeavour misses the point. Who is to say what value particular information may have? As part of the Journalism and Media Studies Centre at The University of Hong Kong, I know that the value of information is sometimes only seen over time and after much examination. The true benefits of open data come when a government believes in the importance of exposing how they function. This leads to greater scrutiny, but also greater understanding. The well-functioning government works with all its stakeholders, and this includes its citizens and public and private organisations.

The Code on Access to Information sits at the heart of the role of open data in Hong Kong. Many have been calling for a Freedom of Information Act to provide a legal framework in which to reliably get information. Even setting this issue aside and focusing on the relationship between a government and its information, it is safe to say that the Government is in the information business. It is essential to its health to produce as much public information as possible The Hong Kong Government has a robust press release system, and yet it suffers from two main issues. The first is that press releases are in a non-machine-readable format and there are not sufficient mechanisms to programmatically search the archives. This is a small technical issue compared to the second issue. The real problem is that HKG (and they are not alone in this) still believe their role is to explain the meaning and value of the information released. This process is an archaic view of government information. Just like Data.One’s view of ‘high-valued’ datasets, the value of information should be open to the experts in those fields. One of the many jobs that Journalists provide is to make sense of information. When we to work with raw data, we are able to connect the dots in a much more meaningful way. Instead, we are faced with decrypting a typical Government press release. There is truth in data. The Government of Hong Kong does good work, and it behoves itself to illustrate that with its data.

As a more practical suggestion for the passage below (#34), I would encourage the OGCIO to incorporate a few ideas:

The first being that HKG needs to create a ‘culture of open’. This is perhaps a slightly strange term, but it truly embodies the intent that governments the world over should have. And since the 2014 Digital Strategy is meant to be a roadmap for the next many years, it will be essential that it articulates the principles of using technology to improve its information architecture and exchange. And this is, it should be noted, not specifically a technological issue. Although technology appears to drive much of what we do, it is nevertheless grounded in deeper principles. If the section on open data does not include reference to opening and improving the access to information, by establishing a long term goal of releasing information ‘early and often’, this passage will do nothing more than position the government as a non-agile organisation.

The final point is that government departments will look to this strategy as a way to interpret their relationship to information. The more open the language is, the more encouraging the OGCIO can be. We have found in many cases that some departments are simply unsure whether they can release information when it is requested. It is somewhat telling that the Guide on the Code of Access to Information is almost 5 times as long as the actual Code. Hong Kong should learn from the United Kingdom which has set in motion a well-established FOIA, which includes a very meaningful archives law. They have also gone on to guide their employees with things like the need to define a public task. This kind of guidance will help to establish a culture of open and enable quicker innovation and advancement in technology, in a way that a prescription for more machine-readable data (and a few APIs) could ever do.

Thank you for the opportunity to provide some feedback. I look forward to seeing Hong Kong embrace its role as a 21st century global city.




34. Through different channels (like press releases, publications, websites, etc.), the Government releases a lot of information in different areas. However, most of such information can only be read but cannot be used. In view of the immense benefits of widening access to PSI for free and easy re-use, we propose to make all Government information released for public consumption machine-readable by default. Where appropriate, datasets will be released with application programming interfaces (APIs), providing predefined functions to make their retrieval easier.

Why is Open Data difficult for Governments?

I’ve been asked this question a lot lately. And recently, a journalist asked me this during an interview for a story on his problem getting data on noise pollution from the Hong Kong Government. Christopher DeWolfe quoted me in the article, Hong Kong’s Silence on Noise Pollution (which was also printed in the SCMP and is online behind their paywall):

Without open-source data, there is no way to see if a similar situation exists in Hong Kong. “There’s potential for the government to be embarrassed – it’s like having an auditor, and that’s why they drag their feet on releasing information,” said Darcy Wade Christ, a researcher for the University of Hong Kong’s Journalism and Media Studies Centre’s Open Government Data initiative, which is pushing for more transparency in the government.

This question comes to me a lot because of my work at the JMSC on the OpenGov project, but also because I am very involved with the Open Data Hong Kong group. I am more than happy to discuss this issue because I think it is at the heart of how we begin to form a new relationship with our governments all over the world.

As I said in my quotation, opening data has the potential to cause a lot of trouble for the government. This explains, in part, why they are reluctant to open up certain datasets. There are obviously other reasons, including costs required to convert or manage ongoing public data. But by far, the most important is that Governments do not know what opening certain data will do. All governments will be conservative when it comes to things they do not understand. Technology (especially the Internet) has always presented this problem, because it is new, always changing and its effects are unknown. And since there is no government without politics, all decisions pay some heed to the political ramifications of doing anything.

When I was talking to Christopher DeWolf, I wanted to underscore the idea that Open Data does pose the potential for embarrassment because it enables citizens (and politically motivated groups) to audit and present displeasing results. But it also has the potential to improve how the government functions for the very same reason, in that it provides additional auditing. It can also improve the relationship between citizens and their government by engendering more trust. For these two reasons alone, it is in all of our interests to continue to find more ways to open up public data. Some people like to stress that any data the government has generated should be public because it was funded by taxpayers. I don’t usually use this argument, because it can end up dead-ending around the issue that not everything should be public – that the government is given the responsibility to protect all citizens, and that some data may negatively affect a certain group of people. I believe the Hong Kong Government takes this issue quite seriously. They are always conservative when it comes to embarrassment. This may have been why I chose to use that word (wisely or not). The HKSAR has a history of defamation lawsuits, but it is also part of a culture that believes strongly in not publicly embarrassing others. There is a time and place for showing your power, I have come to understand.

So when we talk about the reasons why any government should open more datasets, let’s keep in mind that this is a complex issue and that all arguments have many sides to them. The recent global appeal to Government Transparency and Open Data is good, but it will take time. We have to be certain of discussing both the good and the bad. We have to appreciate that governments will drag their feet and given their very jobs (not as public servants, but as politicians), we have to at least acknowledge why they may not embrace the ideals we believe in. But we can still make a little noise here and there. We can still improve how we speak and use powerful words like audit (are you afraid of an audit?). And we can also continue to explain why we want certain data. The Open Data Hong Kong group is often discussing Open Data as a general mandate. In my work at the OpenGov project, we are usually going after lesser understood and known datasets (land use and corporation information) without fully understanding the goal of what we want. We can extol the importance of having more information. We (especially journalists) can begin to articulate a premise we might have – that having certain data will help to prove or disprove our idea. But no matter what any of us do, this process will take a long time, because politically, there is very little incentive to open oneself up to critique.

We could all afford a little auditing in our lives, but who invites an auditor into their house?

Working with OpenData at the JMSC

jmsc-hku-logoAs a consultant at the Journalism and Media Studies Centre (JSMC) at HKU, I’ve been able to see some of the interesting work being done with Open Data here in Hong Kong.

There have been several projects, but most of the work has fallen under two main projects:

The Data Journalism Lab is being guided by the growing need for data (or computational) journalism. Courses have been taught by leaders in the field like Jonathan Stray, who teaches at Columbia’s Journalism school and works for the AP and by Irene Jay Liu, leader and editor for ConnectedChina at Thomson Reuters.

The OpenGov Project is a newer initiative to focus exclusively on finding ways to access and use Hong Kong government data. At this early stage, we are mostly involved in exploring the landscape of data being released by the HK Government (ie., as well as experimenting with what data would be useful, if it were available.

We are also beginning work on a Hong Kong Transparency Report, which will document the state of transparency of government user data requests and content removal requests to information and communications technology (ICT) companies.

What’s next for Open Government?

Anil Dash has a great post on the Knight Foundation blog called the end of the beginning: lessons from open government so far. In the post, he outlines that we’ve done a lot of thinking about what open government could mean, but that we haven’t yet really got started, solving real problems and building real tools. He sums up what we have learned with the following list:

  • It’s not about the technology: While technology is the enabler, changing the culture of agencies, institutions and society at large will always be more challenging than creating or deploying technology.

  • Shipping beats wishing: Though it’s not about the technology fundamentally, there does need to be some set of tools for people to work from and iterate on, and often times it makes sense to start from work that others have already done, rather than hoping for optimal solutions.

  • Things are “broken” for a reason: Everyone can point out countless examples of inefficiency, ineffectiveness or incompetence in civic institutions when looking for problems to fix. But many of the biggest challenges in Open Government come from insufficient empathy for the circumstances that caused those problems. Sometimes there’s a good reason that an agency or office is doing things that seem ridiculous.

  • Goals like “transparency” are necessary but not sufficient: Of course we want visibility into the workings of our government, or access to data that we’ve paid for as taxpayers. But transparency is not a goal that most normal people think of in the course of a normal day.  Projects that merely reveal that money corrupts policy or that some agencies waste funds don’t always galvanize change. In fact, they often reinforce people’s cynicism. It’s better we give citizens way to act on systemic problems, rather than describing those problems more vividly.

Being a technologist who often finds that technology is raised up as a cure-all, I’m pleased to see the de-emphasis on technology. So often we believe that technology will solve a problem, when in fact, we have yet to identify the problem. With the release of Evgeny Morozov’s new book, To Save Everything, Click Here: The Folly of Technological Solutionism, we’re beginning to see a counter argument formed. Namely, that sometimes our interests in using technology actually create new problems and sometimes don’t even properly solve the one we were working on. Open Government initiatives are very ripe for this kind of problem.

The second point I really agree with is the idea that things are broken for a reason. Part of trying to identify problems and solve them should force us to deeply understand the system we are trying to improve. I recently came across a nice (but old) idea: Le Chatelier’s Principle states that the system tends to oppose it’s own proper function. I think that Governments produce some of the world’s most complex systems. They are created by people (so clearly they don’t have Nature’s simplistic design sensibility) and it is continuously changed and tweaked, often without removing broken parts or inefficient workings. Of course, we would like to believe that they do in fact evolve, but part of what we may be missing is that they are still lacking key inputs. For me, this is the role of Open Government. So, although I fully agree that we do not do enough to understand an existing agency or office and that the way it functions may very well be for a reason, I think that part of the what Open Government initiatives can help us do is understand why an existing organization functions as it does and look towards ways in which our role, as a new input, begins to affect the system. So first we start with empathy and understanding and then we move towards our active engagement in the system.

Of course, at a certain point, we too become part of the broken system. We can’t escape entropy, especially the entropy of information.