 I decided to try to share some thoughts on open data and open source from a government or a public agency perspective. I don't think this is an audience that needs to be convinced of the value of open source, am I right? How many people believe in it, hands up? Okay everyone, okay, don't need that. But in order to make it somewhat interesting, I thought I'll share and review a biased view from someone who's usually on the opposite side of the fence from you guys. Well, one and a half years ago, Singapore was blanketed by hate. Basically, some not-so-good chaps lit some fires in the forests of Sumatra, Riau in particular. And the problem with the forest there is that it's actually growing on peatland. For those of you who don't know, peatland basically is carbon laid down by forests that have been there millions of years ago. When you light a fire in that kind of forest, it's not just the trees that burn, the ground literally burns underneath you. And it burns for kilometers subterranean and then surprise, surprise, you get lots of caves. And it's basically small particles of carbon and other organic materials. Anyway, Singapore is blanketed by that and the worst occurred on the 21st of June, 2013. This is just a photograph of an air quality sensor and we've got several of these scattered across the island. And in particular five of what we call reference stations that supply life feeds into our system. Now, this is some of the stuff which that's not so interesting. How many of you know what drums are? Anyone? What are the drums? No, it's not a filter. Yes, it's a musical instrument. What else? What else the drums mean? Distortions, rumors, untruths, misinformation and smears. So you've learned one new algorithm today. Not my invention. But let me explain to you what happened on the 19th of June. The case levels suddenly sparked. I knew about it because my dinner was interrupted and I had to rush back to the ministry. What you see here, however, is a screen graph that was circulating widely a few hours after this. And what I wanted to show you, you see that circle there that will zoom up. How many of you can see the figure at the bottom right-hand level that says 393? Can you see it? Okay. Actually, the number that we recorded was 321. How did 321 become 393? And how did this manage to convince so many Singaporeans that we're trying to pull a fast one? First of all, you guys are familiar with Firefox, right? I mean, I just saw it. You're familiar with Firefox Developer Edition, right? You know how to download a page, right? And after you download a page, you know how to edit it, right? It's quite easy to change 321 to 393. And then you view it, and then you print screen it, and then you circulate it, and you say, aha, caught you. It was actually high, but you subsequently doctored it to 321. And during times of crisis, it's not facts that travel. It's scandal. It's distortion. It's rumors. It's misinformation. It's smear. So the next day, I had to deal with a problem at multiple levels. First, what to do with the hay. Secondly, how to convince my neighbors to put out those fires. Thirdly, the whole issue of credibility and trust. So following on my earlier experiment trains, I realized we now live in a time of age when governments cannot resolve to censorship. You can't shut down internet just because someone is publishing a falsehood that is inconvenient. And it reinforced my belief that you need to move into open data because you have to demonstrate and prove transparency in order to gain credibility. And in order to do that, it's not only enough to just say you're transparent. You've got to show it. It's got to be real-time, and it's got to be verified. Which means, in the example I showed you earlier, it would have been very helpful if someone, somewhere else, had at that same moment that same data and would be able to prove instantly this false doctrine. So that's the first part. Really about credibility. Now data.gov.sg is a government website which has actually been around for a couple of years. We had about 8,000 data sets on that. I was told yesterday they've just added another 5,000, so it should be 13,000. But I haven't verified it yet. In the spirit of openness, go and verify that for yourself. But I wanted to highlight that of that, there are 118 data sets like my ministry related to the environment. Now, the problem, when you go to a typical service and say, I want the data, it's like, yeah, sure, put out a PDF file. If you're lucky, you get a CSV file. If you're a bit luckier, maybe a text file and Excel, especially. But a little less than the prime minister said, no, that's not... When you put out data, I want it to be machine readable. Because even back in the days of the haze, I realized people are writing iton scripts to scrape my website in order to get the data. And actually, we should have just provided the data. Machine readable, accessible in real time. Everyone could look at it. Everyone could verify and move on through that. So what I've done now is insist, at least for those guys who report to me, that I want APIs. I want application, what is it, application, programmable interfaces. And to test it, I want to be able to just write the URL and see something spit back at me. I tend to keep my civil servants busy. So about a month ago, they said, okay, we've done eight. And this is some... So if you actually go to this website now, www.nea.gov.sg, Obleed API, you will see these datasets. You will of course have to register. There's just a basic precaution. But if you do that and you put in the appropriate URL, you will get this XML file. Because they sent it to me at 11, I decided to test it. And it just so happened at this point in time, I'd been trying to learn a new language. I decided to pick up Java script. And then I decided, well... Is that an unbuttoned language or not? And then they told me, well, go and look at Node. Node.js. I said, well, I stumbled on Node.red. And this was also because at that time I've been fiddling around with adrenals, which actually Harish first introduced to me several years ago. And I'm not sure how many of you are familiar with it, but Node.red makes it trivial to write programs that link objects and APIs with some simple logic. And I thought today I'll just show you that as an example. Because the main point I want to make is how... So let me just get to browser. And this is what it... This is all... What happened to that? It doesn't like my safari. It's amazing that we can do a lot of things with the simplest technologies to get us. Okay. All I have to do, basically, if you think about it, I've already got it. I've registered. I've got an XML file. Basically what I want to do, take that XML file, convert it to a JSON object, decompose the JSON object, and be able to present real-time hourly concentrations of PM2.5. And to make that available on a website. And with this framework, it really is trivial. So as you can see, all you see are just boxes with lines. So the first one basically allows me to write a RESTful API called Air Quality. The next one, PM2.5, I think you can see it. All it does, it just calls. It just calls the API. Next step, translate the XML file into a JSON object. I didn't even have to type any code at all. Now, the next step involves a little bit of coding. But as you can see, it's not really rocket science. Basically, it's some standard, I think it's standard. I mean, I don't know, I'm self-taught. It basically allowed me to decompose my JSON object to pick the attributes that I wanted. So what that does after that, and then I have a template so that it looks, I can control its presentation layer. And then, when I move on to the host, and I now click it, it runs a server for myself. And if I wanted to make this available, I'd run it on Bluemix or something like that. I don't really understand why you need to run servers anymore nowadays, but never mind. And you would get that. But that's not all that it does because I also get it to send email notifications. So for instance, here, now, I'm going to change this email address to Roland. That's it. Okay. I should put it under, I did not know it was going to do this. It was really interesting. So what this does is, apart from providing a website where you can access it, it's also sending the same information to the email. So hit Deploy, then Refresh, and sometime, hopefully soon, if I'm assuming I've typed it correctly, Roland should receive it. But look at not just emails, but look at all on the right-hand side. Can you see all the other stuff which I can do? I can send a tweet. I can put up IRC's. I can tweet email IRC. And so if Roland receives a lot of spam from my email, it just means I've done a error in my programming. Just one. Okay, so I didn't prove it. But anyway, so this was a little test between 11 p.m. to 1 a.m. The purpose of this was really first to test the APIs they're giving me work. Secondly, to make the point on how trivial this whole exercise could be. Can I get back to my... Oh, will I lose it? No, I want to get back to my... Okay, I think I'm okay. Recognize this chat. That's actually a serious message behind that. In the midst of dealing with the problem, Newspapers reported this guy, Roland Turner, who doesn't believe your results and is creating his own senses. Okay, I am as I break it. Now, think about it. From my perspective, I've just said, now he's completely wrong. He's using a different methodology to measure particles. We're using meter waves. You were counting particles. We're using meter radiation to measure weight. But I realize a far more productive relationship was to recruit him, put him in touch with my chief scientist, get mutual education done, and to make him part of the selection. So really the main point of my talk today is that I'm a believer in open data and open source, not for ideological reasons, but for very practical reasons. It enhances transparency. It enforces integrity and honesty. It creates opportunities for collective ownership of problems. So even if you have a different opinion from me, a different method from me, but we can still hold the problem and hopefully be empowered to co-create solutions to the problem. And that opens up the whole ground of innovation and ultimately, entrepreneurship. So for us, and I say for us, it's not just me in government, rest assured that we will support your efforts at promoting open data and open source. We will continue to expand the data.gov.sg site. We will continue to promote hackathons and make that data available to you because I am sure you'll be able to do far more than what we can do. And because you bring a different perspective, the ability to synthesize different views, different data sets, and create novel selections will make a different self-society. So that's basically what I wanted to say. So thank you all, rest assured that you will always have our support. Thank you. So the question I would like to pose is what do you think can be done better to have government agencies have a more, have a quicker approach to implement APIs in the agencies? Because as I see it, this API talk has been around for quite a number of years and there's not much progress, but in fact these are pretty trivial implementations which can be done quickly if you have the right people in place. So my question is what do you think can be done better to have all this data available in a far more efficient and faster way? You're absolutely right. The data actually is already there. The real hurdle is mindset. It's, you know, for most people, especially in the past, data was power. And if I've got it and you don't, I've got power. The mindset change now is to realize that real empowerment and real solutions lie in sharing that data and in allowing different minds to work on it. And that, you know, this is an age in which trust in government is always a challenge. It's not just in saying all over the world. But I think governments which get it and governments which then open up data and open up solution spaces to the community will build a new relationship. So we are doing this, you know, like I said, I'm not doing it for ideological reasons. I'm doing it for very practical reasons. I think it will solve real world problems. I think it will enhance the ability for government to provide services to citizens. It will give citizens a deeper sense of engagement and ownership and bring us up to a new thing. So what we're doing, well, the chief advocate for this is the prime minister. So that helps. If anyone says no, I just bring them up. Explain to PMY, it can't be that. Now, having said that, there will be some data sets which are not going to be open. For instance, income tax data, that has to be safer. Similarly, if there are issues involving national security, that has to be safeguarded. Now, there will be grey areas. For instance, many people have told us, you know, why does the government keep making me fill up forms, name, address, NRIC number, again and again and again, right? Now, on one level, you can say, well, if you share that across agencies, actually that will require some policy and maybe some legislative changes because the way we are structured right now is that if you give data to one agency, that agency is obliged to protect the data and not... So we will have to look at appropriate models, governance models and regulations so that we can protect privacy because I think the big disaster waiting to happen is the identity theft and loss of privacy. I think we're taking it for granted right now. But until and unless you can secure your identity and secure your privacy, otherwise, by the greatest worries, there will have one big scandal and they will set this entire open data movement back many, many years. So I'm appealing to you to understand that we actually are on the same side but we have to do it carefully or risk setting us our movement back. So it's not going to be for lack of trying. Now, the way it works is that some of you in this room, I mean, you've got my email address. If you actually come across a need for a certain data set and you think it's been unreasonably withheld, just drop me a line. We'll look at it. Do you have something you're chasing in particular? No, I mean, I think you answered the question by saying it's a mindset issue and governance model problem which you need to resolve before you can move on. But do you agree with me? We also need to put some bandwidth into this issue of identity and privacy. Actually, I think security is going to be an increasing problem. So in fact, for those of you who are working in these fields of identity, security and privacy, you've got it made. It's a growth thing. So it turns out there were really questions like that. So please go ahead. Good afternoon this time. I came off the media. That didn't show up. So my question is simple. The data sets that you've just released, it's incredible. Duolemen like me, the data... So what is some of the uses you think the community is going to have for these data sets that the government has released, is machine-readable data sets? Well, I think the first thing I should say is I probably don't know what you are thinking of or what you'll be able to create. But my approach is we know certain data sets at times. So we'll keep that in mind. For everything else, the default should be to make it as open as possible. And the other thing is there's a difference on near-real-time data. For instance, the stuff that I showed you is hourly. And data which is released monthly or annually or data series. And this is something I'm still discussing with my... with my data guys. I mean, they tell me it doesn't make sense to release, push out XML for data series. I don't know enough to be able to answer that question. Go for it. So, but again, we will make... That's why I think the best thing is to assume that for most things, the default should open up more and for specific requests, we will consider on its own merits. Thank you, Mr. Well, what are some kind of things you would like to see done with this data? Okay, let's think of the problems that we face. Let's get real, right? What are the some of the problems that we face? Let me give you an example. Transport. Wait back there in Singapore, right? We've got the highest prices of cars in the world. Too many people spend too much time stuck in jams or in crowded trainings getting damn angry and voting against us. Real problem. No, already elections not so soon, so not. But is there a way to use data to make that daily commute more efficient, more entertaining, more enriching? And the answer is yes. For instance, if I told you that we knew the location of every single bus and every single taxi in real time, would this be relevant to planning more optimal routes? Would this be relevant to having express services? Because you know, for instance, how many hundred people are queuing up at that bus stop? And you know where these people are going. Can you devise? Rather than someone just sitting behind a desk and roughly gestinating with, what if all these data are available in real time? And not trapped behind proprietary dogs because we know companies all want to build firewalls. They want their wall in gardens. What if all this was available out there? Now at the same time, I've got to make sure you can't chase your ex-girlfriend location in real time. So there are issues of privacy again. So this is some example. I don't believe we can make transport better. Environment, apart from getting, this is just a simple issue of transport. Apart from that, what issue, for instance, is literary? Why are some people, a minority, persistently literary? What is the pattern? Who is doing it? When is it happening? Should our cleaning and our enforcement be more targeted, more directed? So there are a lot of mundane examples. Well, not so mundane examples. Healthcare. Why is it, if you go to a different doctor or a different hospital, sometimes tests get repeated? Why? Now, does your healthcare data belong to you or to Apple or to your doctor or your hospital? And the answer is there. Actually, it should belong to you. Now that's easy to say in theory, but we've got a problem. Until we can secure identity, we can't have an electronic medical record that is accessible to the web. Unless anyone of you has solved this problem, but we can't. So, you've got a solution. No, not a solution. One question was... Yeah. So the point is healthcare can be made more efficient, more accessible and hopefully more cost effective if data and the latest information was brought to bear and made available to both the patient and the doctor. So again, this is another example. So we are trying, but I hope you see my point that the limiting factor is not politics or ideology. The limiting factor is how can I do all this and protect your identity and protect your right to privacy. So I think these are issues which are going to occupy our minds and I hope you all will be part of this conversation as we evolve. Thank you. So actually to the point about protecting identities, the solution isn't in terms of... Shouldn't we look at that there's no tactical solution to it? From my point of view, it seems like the government do not have the right people to implement the technology because technology is already available and in the States itself, there are multiple startups doing exactly this thing that you're talking about which is empowering consumers to own their own medical data. I myself, I'm working in a startup which focuses on that type and that is one of the problems that is important to consumers and it is not without solution. There are plenty of solutions. I think there are plenty of solutions but I think we should be humble enough to say that we're still not sure which is the final solution and I think they're going to be competing solutions and we should allow that certain healthy competition to occur and hopefully the best product will win in an open source spirit. I'm very specific that someone says I've got the final solution. I'm not saying that someone has a final solution but I'm saying that there are already multiple solutions up in the market and different companies and different individuals are experimenting heavily into it. So instead of looking at it as a let's wait until the technology has matured before we decide to roll it out in government why don't you look at it from the point of view of let's roll out as a small test in a small, in a controlled environment. Yes, that we will do. There are two things we will do. First is that we will define outcomes then we will define the performance that we want. Call for proposals, requests for information, interesting ideas, pilot it, start it on a small scale, see if it works. It doesn't work, go back to drawing more. So we will move away from the approach that we find a consultant. The consultant tells us to do it this way and then we just tender on fixed specifications. We will move towards outcomes and performance classification, requests for proposals and information, pilot trials and then learn and improve it along the way. That's how we're going to do things for now on. I'm so looking forward to that. We are. We are one at a time. One last question. I've been told I can. I'm going to ask you a question. Why not let them all have this. I'll keep quiet for a second. I have a very quick question. Firstly, thank you for coming to share your perspective. My question is, how can we bring your enlightened mindset to other places, countries, cities, that perhaps living in the past. I live in the United States. This is my third day in Singapore the first time I've ever been anywhere in Asia but it's made me realize that a lot of the United States is living in the past. So I would love to hear your advice to how to bring your mindset to other places. I see that your mindset as a politician is very rare and people like us really want to spread our minds. What's the point of that? I'm going to let everyone ask myself because I think too long. I think what he said is very correct. Actually I would also like to see this that your mindset has been transferred to other nations. I have lived in three or four of myself and very rarely I find such examples. So it was great actually seeing somebody really doing something what he breaches others. That's one thing. My question was regarding, when we were talking about it was related to the medical information right. I do understand that there is a lot of talk on the personally identifiable information regarding patient data, insurance data and this one. In my experience while I was working with the pharmaceutical industry I realized that you can anonymize that part of the information but information of the data which can be made available is like what kind of disease numbers which are spreading in those areas. What kind of hospitals with how many beds are available. What kind of doctor number with specialities are available in those areas. Are there any plans to bring back that kind of data into public especially for public medical hospital maybe not private. Thank you very much for the opportunity and for you to come here and speak. Congratulations on taking on the role as the minister for looking after the smart nation. It's specific about smart nation. My concern is again you talked about privacy and security. I haven't heard yet from anyone you have articulated perhaps is what is the basic underlying principle of the smart nation. One of the things I have been looking at is to make sure that there are three principles around it. Three pillars. It is built securely from get go. It is built on open standards from get go and that is an open source reference implementation of whatever is being built so that we can always plug and play from wherever else was provided. Do you have any thoughts around that that's an important aspect if you want to go forward? You don't mind answering the questions but your point, build securely. We believe that security has to be built at the initial by design. Not create a system and then worry about plugging the holes. So yes this is an important pillar. That's also why we've just created a cyber security agency because without security all the other things your other point is open standards. Yes this is something I believe in. I sometimes have to push against resistance but I believe in open standards and as you've understood from my talk the reason I believe in it is because it allows more players more competition and more innovation to be brought to mind and then ultimately there is fruit in better services, better value products and hopefully better business options for our people. The larger question behind smart nation is this. You know the political problem in most parts of the world right now is about inequality and middle class stagnation. Many people I believe are barking up the wrong tree and the political left and the political right are arguing about services. The reason why middle class wages are stagnating is not because this is a right wing conspiracy it's because we are living in the digital age and in the digital age it's not just equality of opportunities but the ability to exploit opportunities that makes the difference and we are now living in an age when if you're the owner you're the most popular singer, actor, sportsman as long as you're number one you are the top of an exponential curve. Number two is at least two thirds or half of where you're at. So in the same way we're getting very asymmetric distribution of the spoils of the digital age. You see this at individual level you see this at company level. The biggest company market cap right now today is Apple, I think it's 600 plus. What's number two? Number three. What do you think? You end up with Google. You know what Google's market cap is? It's only two thirds. So what I'm saying is at individual level at company level you see very steep curves. The other phenomenon which you're seeing you know the business consultants like to talk about S curves and the company grows slowly then you have a rapid expansion phase and then you plateau, right? And those S curves used to take 20, 30, 50 years. Well guess what, today if you look at the S curves of the internet companies you're lucky they're 10 to 20 years. So what all this means is that if you have a country or a city or a people that are not prepared for this there's going to be cut branching changes there's going to be unemployment there's going to be the depression of middle class wages because routine white collar middle class jobs are going to be replaced by machines and smart technologies many of which you all are going to create. So when Singapore says smart nation it's not another fan or another ideological thing it's actually born out of a bit of paranoia that we need to stay on the right side of the curve. The reason why we're spending we're going to spend a billion dollars a year on skills future. The reason why we're going to try to teach all children to code or at least understand computational thinking the reason why we want to make data available to everyone so that you can massage it you can analyze it you can create new products is because we want to make sure Singaporeans are not fighting a losing battle to be cheaper, better, faster than a robot. We want Singaporeans to be designers programmers operators of the robots not the losing competitors of the robot so actually that's where the real energy or the real the real angst behind what we're doing so yes there is a sense of urgency and I agree with the principles that you have been unseated now the third point you made was about open source open source implement reference implement now this part I'm not going to be excessively dogmatic I think where possible and where it makes sense it fits for purpose we should do so but let me tell you one reservation I have with open source which is that as I look around the room I wonder how many of you are going to be millionaires it sounds cool to talk about it but actually I do need at least some of you to be millionaires on the basis of your open source work because I volunteer but it needs to happen because otherwise you guys are all going to work very hard otherwise it's going to make the profit right? and that is not the basis for a sustainable movement so I'm just sharing with you my anxiety about the open source movement whether it will make enough millionaires of the people like you who are going to work so hard and spend so much many years of your life chasing I'm not going to be excessively dogmatic if you finally come to me with a product and as I said in the way we're going to do it is we specify outcomes you come to me with a product you prove that it works I will of course ask you whether it can be open source and you say no it can't and you can convince me I'm not going to rule you out simply because open source but yes I know many of you will still make a healthy living a good living by being consultants but just bear this point in mind that look around the room if we can get how many millionaires do you want down here if we get 100% our problems are over but it's not going to be so simple so that's my my slight anxiety about this point on the other two questions the question on data you're right if for instance data on waiting times data on heatmax of crowds can be suitably anonymized and lead to better delivery of services we will do so but very high there was this study I think they were doing it of New York caps it was supposed to be anonymized but with intelligent analysis they were able to identify who were the muslim cap drivers based on their patterns like times of day or particularly on Friday so bear in mind that even with the best of efforts there are still some risks and I hope you agree with me where people's financial and health data is concerned we have to lean more towards security and privacy rather than experiment so by all means experiment in my environmental data I don't think there's so much risk to your privacy and I'm in the case of the other advantage the Singapore government has and this relates to the first set of questions about going to other countries we've got a couple of advantages our prime minister is a mathematician and was coding decades ago half our cabinet are engineers or former engineers or if they're not engineers then they're surgeons or doctors like me we understand science we understand mathematics and statistics so we sort of get it I mean we may not agree with us with our political views but I think on science and on data and on programming we get it it's quite different in other countries where the majority of politicians have spent their entire life thinking they've never had a real job they've never programmed something they've never run a company it's very difficult so no we can't just go out and say be like Singapore and in any case our final and biggest advantage is that we are small it's much easier to run a small place I can sit down here and talk it's very hard to do in a much bigger country so the level of the decrease of separation in Singapore because we are small is much less than 6 so there are things that we can do here which is much harder than any other place so I'll give a final example we are not Silicon Valley we will not even pretend to be Silicon Valley Silicon Valley is unique it is where the transistor was invented there was that unique combination of academia and finance and business together with the internet and the Department of Defense Investments I think it's a once in a lifetime place where not only do you have technology not only do you have money you've got smart money money that angel investors were able to mentor, able to network we don't have that debt but what we can offer is I was in Silicon Valley a few weeks ago and just asked about bandwidth and there are many places in Silicon Valley where even if you wanted to pay for it you can't get a fraction of the speed of the weekend we can deliver fine but yes we can argue about how slow sometimes when we say slow it means a couple of weeks we can argue but we can deliver it to every home so because we are small because we have a single layout of government on the infrastructure side of you I intend to be top of the world there's no reason for us not to then having done that we can then open up our data open up our problems open up our challenges worldwide if you've got a solution that will solve health care transport whatever real world problem proof of concept is in Singapore it works in Singapore we up scale it to a national level you will have opportunities all over the world because your calling card will say I solved it in Singapore come to Singapore see my product in action now let's start talking about how it will work in countries which are far bigger far more complex than Singapore in other words we just we cannot be the centre of the universe but we can be part of a global node of smart nations and smart cities part of that value chain as long as we can capture some of it for our people and for the friends of Singaporeans I think we'll be alright so I hope I've given you some idea of why we're doing what we're doing and some idea of our strategy and I also hopefully convince you that we're going about it in a deliberate, careful and reasonable way without hungry and without arrogance and that it's a realistic plan and finally as I I just want to emphasise again I really want all of you to succeed and because I think this goes far beyond just a preference or an ideology I think many of the challenges that the world's confronting can be solved with the values that underline the open source, open data movement because it is based on honesty it is based on rigor it is based on sharing it's based on innovation if we can do this and replicate it we'll solve a lot of world problems so thank you all again very much I really, really wish you all the very best thank you