 But nonetheless, we have Mr. Emerson who will be joining us and telling us today about this. Hi, everyone. Yeah, so, please proceed. Okay, good afternoon. Wow. Okay, so I hope your Congress has been good so far and you're not too hungover or too tired. So today I am talking along with my colleagues from Sierra Leone if Skype works. If the Skype... Yes. Great. Yes. Awesome. All right, guys. Make some noise. Let them know that you're here. All right. Awesome. So, great. So, today a talk is about what, by most standards, would be considered a miracle because it... So, a little bit of an... So, here's the structure of the talk. So, I'm going to give like a two-minute introduction. There is a bit of an outline. You can see the structure of the talk and then we're just going to crack into it. So, introduction. My name is Emerson Tan for six months between October 2014 and April 2015. I was the country coordinator for a consortium of NGOs called NetHope. And NetHope is a consortium of 33 of the world's largest NGOs which specializes... And the consortium specializes in technology and IT and development. And I was dispatched to Sierra Leone to be the country coordinator with the world's vaguest orders because they could not find anyone who wanted to go. So, I got three people's salaries and they just said, go there and improve things. And then the orders just stopped. So, that's what I tried to do. These guys... So, Salton, can you wave? Yeah. Okay, that's Salton. Harold. That's Harold. Francis. Yo. Okay. So, these guys are from IDT Labs. And these guys are based in Freetown in Sierra Leone. You can see up there. And these guys are all self-taught programmers, self-taught hackers. And these guys basically saved their country from complete collapse. And I can't possibly begin to describe how many lives were saved as a result of their work. And indeed, we could have lost the whole of West Africa. Anyway. And this is basically their story. So, a bit of an introduction. This is Sierra Leone. Sierra Leone is roughly the size of Ireland. It's on the west coast of West Africa. It's got about six and a half million people thereabouts. Nobody really actually knows because, you know, the census is kind of rubbish. Yeah, right. It's the sixth poorest country on earth officially. And that was before Ebola. The Ebola epidemic. It had a civil war between 1991 and 2002. And that is the source of so many movies like, you know, Blood Diamond. And, you know, you get people like Leonardo DiCaprio running around going like, oh my God, look at those Blood Diamonds and people in Civil War and, you know, all this sort of madness. And that's before Ebola, that's what Sierra Leone was really famous for. It was famous for a grotesque civil war and being very, very poor. But it's really rich in diamonds and gold and minerals and the rest of it, which is why everyone likes to fight over the place. Yeah, it's kind of shit. Anyway. Yeah, yeah. It also has only 150 metres of fibre optic cable in the entire country. And it goes from the cable landing station up to this sort of like microwave tower. And then the rest of the country is all point-to-point microwave. So if the Skype connection goes down, it probably will because it started raining or something. Right. So I could bang on for hours about Sierra Leone and Ebola, but hopefully a documentary crew came and followed me around. So far we are in Sierra Leone doing an Ebola vaccination. Let a part of the science and symptoms bleed from all openings. Bleeding from the ears, the eyes red, that is the end part. When you see an Ebola patient suffering from those conditions, obviously there is no hope. Since Ebola came, schools have just been stopped. Business is not going normal again like the way it was. Ebola has overtaken the whole nation and so students can't come because of the risk. Our country is suffering. We don't know the way our country is going. We don't know the trend of our nation. We don't know what we are heading to. Our children are suffering. No education. They have been locked up. No movement. Everything is going backward. The 80 years break between the tsunami, all these things raised a billion and a half dollars, but we don't see the same public health right here. Within one week, the whole household affected. All the money from it now. Within one day, 92 million people died. We are worried. We are worried about them. Everything is going to be fine. The people are not going to be able to go to school. They don't have the money to go to school. They are going to be able to go to school. They are going to be able to go to school. Ebola will be defeated in Sierra Leone. But in the meantime, people are suffering. We saw some examples last month where people couldn't go to the farms at the edge of the village. But because the village was under quarantine, those people then had to sit there for three weeks, relying on food aid, not being able to collect the rice harvest, which then is clearly going to have not been an impact down the road. We came to Sierra Leone because we wanted to tell the story. There's people who are on the front lines. There's so many heroes, so many people who are working to battle this disease, this outbreak, and with honestly very little help. And this is a war. Those in my friend does not mean that I must not partake in the war. So it's my duty to partake. Because I didn't volunteer to help my people. Who would help them? That's the question I've been asking myself. That's why I decided to come and render the little care I can to my people. My name is Asiyatu Emansari. I'm an Ebola fighter. My name is Lieutenant Bah. I'm an Ebola fighter. Hi, my name is Joe and I'm an Ebola fighter. I am an Ebola fighter. I am an Ebola fighter. I'm Daniel Mukunda. I'm an Ebola fighter. My name is Dr. Vincent Ibatwala. My name is Ian Samwell. My name is Aminata Eskabia. My name is Mayatu Frashez Bangua. I'm an Ebola fighter. I want to ask the world to see Sierra Leoneans and even Liberals and the Neons who have had this crisis to see us as human beings. We're all the same. We should not be stigmatized. So, that little trailer does a much better job than I ever could of summarizing it's a war, you know. And in wars, people die. And what happened was that the disease ripped its way through the already very fragile health system. Some of the first people to start dying in significant numbers were nurses and doctors and medical staff. And the system just started to collapse. The country came, all activity in the country essentially came to a halt. Commerce stopped, schools stopped, everything stopped. And it looked like there for a while that it would rip through the population of the country and the UN finally declared an international health emergency, set up a special UN mission. And then you saw all of these countries like the US and the UK start to mobilize their militaries to actually start to help Liberia and Sierra Leone and Guinea start to respond to the crisis. And I arrived in October on an aircraft with some of the real first wave of experienced humanitarian and medical workers from outside. I don't know, there should be a fellow here from eHealth Africa. I don't know if he's here. There he is. Yeah, so I was the only technologist on the plane but it turns out I was wrong. It turns out that Kristen from eHealth Africa was on the plane and I have to salute this man up here if you can't see he's behind you for being brave, stand up. For being brave enough to actually get on the plane because back in August and September, essentially if you went, if you were considering going, people, everyone thought that you were going to die. I went to a friend's wedding the week before I left and I got talking to the priest and the priest was like, oh, well, what do you do? Yeah, I'm going to Sierra Leone. I'm doing this, that and the other. He just stops and then he just gave me the blessing. Like, hand comes out, a bit of holy oil on the forehead and the rest of it. He says something in English and then he mutters something in Latin and it turns out actually that he used to be a battlefield chaplain and he actually gave me the last right. So plainly it didn't work because I'm still here. Anyway, so the UN, the UN WHO come up with this strategy to arrest the spread of the disease. They call it bending the line and what this does is it gets you from a situation where everything's going up exponentially, outbreak, contagion, dust and Hoffman panic, to a situation where actually everything levels out and you get control of the disease and then from that point onwards, the trajectory is downhill. That was 70% of all people who've died of Ebola buried safely and 70% of all life cases in isolation so they can't infect other people in the community. There's only some, that sounds great on paper, because you're like, well, that's easy, because there's only small problems with that because the health system has evaporated and ceased to exist. There are only two major paved highways outside of Freetown. Am I right? I'm right in thinking that, right? Well, it depends on your definition of paved. There's no functioning power grid. There's no functioning running water in most places. The mobile phone network only extends to about 40% of the country. Oh, shit. This is actually a lot hard. So these numbers look great on paper, but it's actually much harder than you think. And we're going to need an army because, you know, at one point, so before the crisis Sierra Leone had one doctor for every 50,000 people. So to give you an idea, Germany per 1,000 people, I think per 1,000 people has 38.9 doctors per 1,000 people here in Germany. They have one doctor for 50,000 people. That's quite a lot. You know, that's quite a waiting line. You know, they are waiting to see the doctor and the rest. It's like, okay, year number 49,578. You know, please don't die before the doctor will see you. Okay, there's some fucking grim humour there. So, you know, it's like, right, okay, so what are we going to do? We're going to need to reconstitute the entire health service from scratch. So it was decided that with the NGOs and the British Army we would have built 18 of these Ebola treatment centres, community care centres, we'd hire thousands of contact tracers, hundreds of safe burial teams, hundreds of ambulance drivers, hundreds of disinfection teams, all the lab staff, and then all of this would have to be coordinated with, like, command centres run by the army, you know, rapid response teams. The 117 emergency line, which eHealth Africa have to give credit for sorting that one out. The lab reporting system, I ended up being deeply involved in rolling out a lot of the infrastructure, like all these V-SAT-based internet to key facilities, thousands of mobile phones, SAT phones, vegan units, sort of this huge infrastructure expansion project, and it all has to happen within about, you know, six to eight weeks, because that's, I sat down, when I first got there I got dragged in to a lot of this largely by accident, and I got dragged into this high-level meeting, because it's like, well, you know, you're the technology guy, you know how to set this stuff up. Yeah, maybe. You know, like, come to this meeting and we'll have a discussion about the way the country is going and what needs to be done. And this very senior, this very senior UN official gives the frankest briefing I've ever heard, when he said, well, basically, we have, like, a couple of months, or we're going to lose the region. We might lose all three countries, and right now there are people in New York who are making contingency plans for millions of deaths. And that's quite sobering and a difficult, you know, so you have to take a minute, center yourself, think through the problems, and try not to get overwhelmed. So, right, I'm going to take a quick break here. Let's see how much time I've got to talk a minute about some of the principles of a successful intervention, because this isn't exactly obvious. So, and these are all backed by little stories. So, be on the ground, not remote. So I went, Christian went. Lots of people didn't. We had loads of offers of help. Google, bless them, you know, IBM, bless them, and others. And their condition was, it's like, yeah, we're really willing to help. This is going to be awesome. We're going to apply all our vast technology, technological prowess to this. But we're going to do it remotely. We're going to do it from London or California, because, you know, like, if I get on a plane, I might die and, you know, and my insurance doesn't cover this shit. So, no. So they totally chickened out. And unfortunately, you cannot develop systems quickly where there are complex business processes and local conditions remotely. You just can't do it. So, those guys, thanks guys. Thanks, but no thanks. Sorry. Make use of local resources and people. These guys understood the environment intimately. They know how everything works or doesn't work. I mean, the classic example is we had to hire data entry clerks. And hiring a data entry clerk here is really pretty easy. You know, you get some CVs and you read the CVs and you go, oh, look, you know, you can read and write and you appear to be vaguely bright. And I'm sure that will hire you. Great. You can type. Sierra Leone, it's a little bit different because most people's CVs are completely meaningless. You know, there are lots and lots of words and, you know, that doesn't actually mean very much. So what these guys did was they came up with an English comprehension and literacy test. It was hilarious. I don't know if you remember that. Especially the bit where the instructions were wrong. I think this is one of Francis. So the instructions on this literacy test were wrong. If you didn't question the instructions, you failed. Yes. APPLAUSE Because what that meant was you couldn't actually read the instructions and understand that they were wrong, or you knew that they were wrong, but you were too timid to actually say, hey, this is wrong. We don't want people like that because they'll be dealing with messy data which will be a mess. So, yeah, 183 people applied. How many did we hire in the end? About 30. 30. So from a pool of 180 people who replied to the advert, we ended up with 13 who could actually read, write, and actually question what they were reading, which is kind of amazing. The real literacy rate, officially the literacy rate in Sierra Leone is 35%. The real literacy rate, i.e. comprehension and being able to actually work it out is under 20%. That's one... fewer than one in five people can read and write an enumerate and a competent. That's shockingly bad compared to Germany where your literacy rate is over 99%. This is a very difficult environment to work in. So, next principle is take chances and fight inertia. Right, so you're not supposed to hire people like these. None of these guys here is over the age of 40. They all use this dodgy open source stuff. Dodgy open source stuff. Yay! And it's like, normally for a huge ERP deployment who do you go to? You go to SAP or Accenture or whichever one's money. What you don't do is you don't go to a bunch of guys who work out of what's effectively a converted garage. But the only difference is SAP might not actually care whether the project succeeds. These guys really care. They care a lot. So, when me and the team that were looking at all these problems started thinking about this, I was like, look, we've got to use local talent. They're the only people who really understand how everything works and they're much more heavily invested in success than anybody else. We have to use the locals. And then we bullied the United Nations development program into paying for it all. They really didn't want to, but we talked the leadership round and the rest of it and then they took ownership of the process and it became a success. So, you know, all for the best. Use open source software, reuse and leverage what exists. This is kind of critical on what this talk is largely about because what you'll see later on is that none of this would be possible without open source software, without open source frameworks, without being able to leverage all of that historic work that has gone into this because you could not possibly develop systems this quickly for such low amounts of money in this sort of environment with this kind of robustness without the existence of this huge open source ecosystem and the huge size of the open source ecosystem and the fact that it covers everything including really boring things like accounting and ERP and logistics management and stuff like this means that there are open source solutions for all of this and what that enables you to do is draw on all of that and be able to spin up things very, very quickly and then plan how you're going to leave because the problem here is that most NGOs most development aid projects don't plan to leave they just show up and they make stuff and the rest of it the locals don't build any of their stuff they don't own any of it and then the funding runs out the white guys fuck off back home job done and then what happens is that their country fills up with garbage because a good example might be you go to the hospitals and what you find are all these piles and piles and piles of old ubiquity little ubiquity nanostations and stuff like this and it's like what is this doing in a pile and it's like well the money run out the satellite link went away and then all this stuff broke down and we just threw it all in a box and that was the military hospital up in Wilkinson there are pictures of me with piles of old ubiquity gear taken apart and me fixing it because it was like well this is stuff we don't have to buy if you plan to leave what that means is that you ensure that you have somebody there who will take ownership of what is built and once they have ownership of it it means that they're responsible for its onward development irrespective of whether they get funding from outside or not so for example the stuff that Salton and these guys have done is you know it's available it can be reused and hopefully it will be reused in the reconstituted Sierra Leone and health service am I right in thinking that they're actually going to do that it's just it's been done now okay well there you go see that we can make it happen but you have to plan to leave you never want to stay there forever and you never want to actually patronize people by just giving them things forever you know they have to learn to stand on their own two feet open source software open source ecosystem allows them to do that so anyway okay so let's have a talk here about how what these guys actually had to do management information system paying people this bit here actually I'm going to get you guys to actually start pitching in on this because I've talked for far too long and you know these guys are probably bored of hearing my voice so some challenges you know so there's some pressure the health system had more or less collapsed we have to hire you know a couple of tens of thousands of nurses and medical staff and contact tracers and all these other guys and it turns out in the old Sierra Leone health system nobody gets paid for months possibly ever and when we got there we were presented with the payment system for the old Sierra Leone medical system I didn't take a photo of it it was three, you know like A4 boxes of copier paper they literally just dumped these boxes of I don't know if you remember the files when they just dumped the paper and it was like there you go that's what we've got and it was a bit of a mess so nobody gets paid and what that does is it forces people to be corrupt it forces people to take money from patients which is extraordinarily dangerous in this context it also undermines people's confidence in the health system basically it means that the whole thing becomes just horribly untrustworthy and because people weren't being paid for months on end they went on strike and then you started seeing scenes like this these are from actually from Liberia press pictures from Liberia but you saw bodies in the streets on one occasion there was a strike and patients weren't being fed so they just broke out like these are you know like Ebola patients and they just broke out and just started running around and looking for food because you know the healthcare workers had gone on strike because they hadn't been paid for six months so you know like strikes, civil disorder riots, the rest of society shut down, if the epidemic doesn't go stop you know like abating this senior official in another meeting says you guys have got six weeks to come up with a working solution to pay everybody or the country's fucked and they actually used that word because that's not normal there's a middle aged guy and it's basically like right either come up with something or we're going to have to think of something else and it's going to be very ugly and the story basically goes that after this these you know during one of these meetings I actually you know they dragged me in and they said well you know is there a technology solution to any of this and I'm like well it sounds like you need an ERP blank looks across a room what is an ERP system well it's an enterprise resource planning system you know it manages payroll and HR and you know logistics and stuff like that great can you build one I can do that I don't know how I've audited them before but I'm not sure it'll be fine look I'll look into it I'll get back to you tomorrow no you get out and you're just like fuck what are you just starting up to right okay let's okay don't panic don't panic I think this one through sit down with the payment team and that led to a very very weird conversation so you know and this gives you an idea of just how difficult this actually is because you know they sat down in these meetings and you're like you know right well we got to pay people and the normal way of paying people in Sierra Leone is with cash and what that means is that if you want to go and buy anything like I don't know fuel you actually need a kilogram of money because that's how I used to pay people I used to pay people using literally kilograms of money and occasionally I would just weigh it because it's easier to weigh a stack of bills that's like this big than it is to actually count them you just go okay well that looks like about three kilos of money there you go you know you know I'll have that laptop here you go and you know this is actually a bit of a problem because when you're talking about trying to pay tens of thousands of people in the rest of it and the banking system has shut down you know we sat down in a meeting with a central bank and the central bank guy is there with along with the World Bank and the RMF and it's like well we're going to run out of bank notes at some point and you're just like what do you mean you're going to run out of bank notes you know it's like well you know we're injecting like 11 million you know dollars into the economy every couple of weeks you know and we're using the army to distribute it all and you know it takes a it takes an entire battalion of soldiers a week to move the money around you know because there's no money recirculating into the system pulling back to the central bank we're going to run out of bank notes you just like what the fuck you're a bank how do you run out of money so so you know like you know you very quickly think of some solutions in your head and you go ATMs right okay well there aren't any ATMs ATMs in the country and nothing outside of Freetown the capital credit cards well there's no payment network for those to work with so that's out how about online banking well fewer than one in fewer fewer than one in a hundred people has a bank account and there's no and like most people can't read and write so no how about accurate staff list can we just sort of work out who we need to pay well no not really because that doesn't exist either because there are only 12 surnames in the country and it turns out that roughly roughly you know like we think that roughly 60% of the staff list from the Ministry of Health actually might be ghosts, duplicate or fraudulent so that doesn't exist oh shit well they got lots of patronage and you know plenty of corruption that's lots to go around and the guys from the World Bank were like you know you have to get you have to put a stop to the corruption of the patches or we're going to stop paying you know and that basically means the end of the country so it's like okay well plainly we have to do something about this so I screwed up making the presentation so some solutions and this is where I'm going to bring these guys in so Larissa are you guys still in touch with Larissa? yeah yeah okay so Larissa was one of my colleagues on the payment on this on this UNDP-led team and Larissa lives in Sierra Leone, has lived there for a very long time and she knew Sultan and IDT Labs from some other previous project and I was describing this meeting where it's like yeah we just need an ERP and you know it'll be fine and she's like you know I know these guys and that's actually what she said, she said I know these local guys and I was like right okay well plainly nobody else is going to shop bring him in and so Sultan showed up Wave Sultan this guy showed up and he showed up with a laptop held together with some sticky tape and he then showed me his code and this is the key moment really because we started talking and we started talking Python and then he shows me his code and I'm like well actually this is pretty good you know structure's good you know it's like I can logically follow it I can actually read this this is actually quite good how long did it take to read this to do this like a week or whatever and it's like this is actually quite well thought through and then he showed me the it turns out they used open ERP before to build the human resources system for a local cell phone provider which is quite small it's got like two well no it's about a thousand something old users but it was really good they implemented features like for example to combat tardiness if you don't log into your workstation the system schedules a meeting with your manager if you don't show up for the meeting of the manager it automatically fires you suddenly Afrocell this little cell phone provider had a 100% attendance rate and what really actually sold to me is that you know because I've been talking on the other phone infrastructure stuff I've been talking to the the CEO of Afrocell and he was like yeah the system keeps firing me automatically because you know because what it would do is it would say you haven't gone to your workstation it would then schedule a meeting with himself because he's the CEO it's like the schedule the system schedules a meeting with your manager which was him and then of course he never bothered to log into it to say that yes he's got a legitimate reason for not showing up to work so the system fired him every week so I'm just like you know what if you can do it for a little mobile phone company what's the difference like 1,000 people 30,000 people it's just scaling isn't it right yeah you know I just have to invent some business process and stuff it'll be fine maybe yeah I'm grossly oversimplifying so you know like that gave me confidence that they could do it and I'm like you know it turns out that I'm also the world's worst motivational speaker because I think I remember in one meeting with Sultan and the gang it was like you know well you know we're going to give you the job of building this thing it's going to have to be done in record you know it's going to be the world's fastest large scale ERP deployment it's going to be the fastest large scale payment system deployment but I believe you can do it principally because if it all goes horribly wrong and the country collapses the British army will pull me out but you guys will have to stay you know it's like I'm a terrible motivational speaker so it turns out that we did discuss in those meetings a number of crazy ideas to try and get money into Sierra Leone at one point we even discussed dollarizing the economy it's like you know they're running out of bank notes right so why don't we just introduce US dollars we'll just make US dollars the currency that'll work well won't it I mean what could go wrong look at Iraq wait a minute alright let's just abandon this idea move swiftly onwards what could possibly go wrong so and that turned and then that discussion turned into well okay what can we use that aren't bank notes that doesn't involve disrupting or totally cratering the economy well what about mobile phone credit because mobile phone credit is pretty much cash right you know if you can turn mobile phone credit okay well why don't we just use mobile money 40% of people here have a phone you know if they can't read an SMS message they can find somebody else who can read an SMS message they can talk to an interactive voice system yeah alright let's make mobile money no problem right and then you find out that in fact actually a lot of what you need doesn't exist because you know like one company had a mobile money scheme Airtel and this is a happy nurse with an Airtel thing but they only covered like 20% of the population so it's like right well we need to get the other companies on board and they don't have a mobile money system at all so we're just gonna have to force them to make a mobile money system with the systems that they have with a presidential decree which I wrote at three o'clock in the morning so I'm just like you know it's 3am they give me policy control okay do do do do do yeah let's make it legal and of course you know if you've seen Star Wars there's this wonderful scene where the emperor is there and he's like you know but so is it legal? I will make it legal so we ended up having to extend the entire system we ended up having to create a digital money ecosystem which is sort of hilarious so and this is the approach that was come up with so it's like okay so we have a comparatively advanced and I use that word very very loosely comparatively advanced mobile phone based digital payment ecosystem yeah that's kind of sort of true at the beginning was very true by the end we've got a good solid local social enterprise these guys who are able to pull it off and we have mobile phones mobile phones are really really cheap it turns out that we bought something like ten thousand little sort of you know cheap ten dollar phones and it's given away like candy you know it's like this is how you get this is your mobile wallet this is how you get your money have a phone and that kind of there were terrible phones they I still hallucinate the ring tone from those horrible little red phones because they were everywhere and they only had one ring tone and it drove me nuts I you know those occasionally I sort of hear them and how to work a miracle so right at this point I'm I know Sultan can you see can you see my slide yes I think so okay please God please Skype God do not say that now okay I sacrifice the vegan this morning for the so that this would work sorry okay so what I'm going to do here is I'm going to describe the problems and then I'm going to get Sultan and these guys here to actually kind of right talk through a bit of the solution so there was no worker ID team we couldn't use fingerprints and biometrics because that was a cross-contamination risk if you were ill and you swiped fingerprint reader and left sweat all over it and then somebody else came along and swiped it you would get Ebola and die not cool most people you know and they weren't really keen on that so we had to come up with a way to bring people biometrically into the scheme because bear in mind only one and five people can read and write and most people don't know what their birthday is and some people don't know how to spell their names and some people might get different people to write their name out and it might come out differently but will be the same guy and then some people share mobile for is a disaster so these guys so we used ODK to make an enrollment application and then we used what was the open source facial recognition framework that you ended up using it's called OpenBR can you describe it so it is built of an excellent computer vision library called OpenCV so what we did was we just wrote a tiny script on top of OpenBR that will in essence compare the feature pictures that was brought back from the field one of the things that we did do is we had great fun in the process of trialling in Sierra Leone this voter registration this voter registration program which we completely wrecked because we took their voter registration machines and dismantled them all to make our health worker enrollment kit yeah I think they're still very pissed off with us for doing that they are yeah we wrecked a few million dollars worth of work whatever so data repository we used OpenERP anyone here works on OpenERP no they're all probably off somewhere the dedupe is quite interesting because the terrible data and the rest of it meant that these guys had to deduplicate everything they implemented this library it works right yeah of course yes it did work the problem was the quality of the data we received was extremely poor we would even have we would have data on the individual's names to supply and we had 27,000 records there were not enough unique IDs things that you can use like states etc numbers were absent for most of the records so that meant we could not use the traditional way of deduplicating deduplicating we could not use maybe SQL so we had to resort to dedupe dedupe is basically a machine of the library and it is written in a pipe it's actually quite interesting and robust piece of work if you're interested in machine learning it's well worth looking at dedupe because it's really, really quick and it basically saved our bacon it is implemented in python so you're not going to get the blazing fast speeds but speed was not too much we were more interested in recall and precision to find the right balance than the speed we had deduplication routine was being run at night so we didn't care too much about how long the result in the morning for an analyst to go through so what we did was if the algorithm would spit out a score and what we did if the score was above a certain threshold the application would go ahead and deactivate basically merge the duplicate sets and if it was between that threshold it was sent into a bin from which an analyst would it was sent into a bin and the difference between the two entities was displayed and an analyst would go through and say is this long and interesting thing is the feedback from the analyst was looped back into the training sets we used for deduplication so it was continuing iterative training communications people can't read and write and stuff so we ended up using SMS dumb phones we ended up with interactive voice recognition and another open source product and corruption is a way of life we just used the did we end up using all the auditing features were the auditing features that were used were those part of the open ERP core feature set well in fact a lot of the the end of the day what we about the only thing we used number one was the core ERP the framework the ability to develop fast the user interface there for you we had to to build a lot of the things on top of the core so I have to speed things up because we're starting to run out of time here so the government demanded that all the data be hosted in Sierra Leone which it turns out is a terrible plan because there's no power grid no data centers, no goods to submit no infrastructure, no nothing and I ended up just bullying them into letting us host it all at Amazon so it turns out that Amazon is actually fantastic for this sort of thing because instantly available they've got all these tools in the rest of it I don't know if my friend Catherine is here from from Amazon good work, keep it up well done Amazon for producing instant infrastructure because without them we'd be fucked deployment management and DevOps the speed of development is actually quite what this system was frightening in the space of under a month how many commits did you guys put in so firstly we had to put together the entire system in just under two weeks and it meant a lot of time at the office it meant a lot of energy a lot of coffee and so on and you know we had to respond to number one how the system was being used what was happening on the ground in the space and number three you know in the least management process that existed so in the space of a month from the time we deployed the system we had to make we had to make over 300 commits 300 commits we had to deploy to push to implement and implement several times a day so yeah frightening pace of development the core system was built in two weeks which if you think about building a huge ERP system in two weeks that's amazing and no one in their right mind should ever try to do that ever again no that software development that's unsafe at any speed it is I still have some health concerns from taking too much caffeine so we could go for like three days without sleep three days of coding without and it worked that's the amazing it turns out that coding for your life really open source software restored faith in the ability of the government and systems to be run honestly you can't put a price on that people were paid on time 100% of the time for the first time in Sierra Leone's entire history but as Francis and Howard will tell you business process development implementation is much harder than technology in these contexts because what these guys did was they organised teams to actually on motorcycles to go and register every single healthcare worker in hundreds of facilities across the country over no paved roads often times in the face of intense resistance there were car crashes people nearly died and these guys here are actually organisational geniuses and they actually deserve at least as much of credit if not more credit certainly me and the rest of the foreigners so please give it up for Harold and Francis and Leslie wherever you are here are some results so you'll see that Guinea programme kind of failed Liberia programme kind of failed Sierra Leone number of Ebola response workers paid on time every time 21,058 total number of workers in the countries the remainder were paid by NGOs 27,000 total that we managed to manage successfully 100% and the one was on digital money that were paid flawlessly 100% we even sent some people to the anti-corruption commission people now know that if there is a system there it works and it is possible to be honest and there are strong incentives to be honest and Sierra Leone owes a debt to these guys for building a system that actually restored people's faith in the ability of that system to deliver and yeah we won really and so here's some good news as of today the Ebola epidemic worldwide is over and that calls for a little dance really because like it and you cannot get up and dance too if you like and that's pretty much the end give it up for these guys I don't know if you see the video and now and and and and and No, no, you're fighting more. The men of three. Yeah, that's you, you'll never see this in the future. They even got the fishmen to dance. And if you've ever worked on any of those open source projects, you owe yourself a huge round of applause because without you guys out here, without your development efforts, none of this is possible. They put in the hard work to save their country. You made it possible. And if you've got any interest in whatsoever, in doing IT in humanitarian aid or development, you can always get hold of me or Sulton and we'll always happen to talk. Thank you. Thank you so much. And that's me. Take a picture of the slide. I'll post, I'll get the slides posted somewhere or other and, you know, you can ask whatever questions you want. Including the Dance of Hope. Including the Dance of Hope. Thank you guys. Also, of course, I mean... We have like five minutes. Yeah, we have five minutes for questions. So please, anyone who has a question, proceed to one of the mics on the right or the left side. So... Five minutes. Go. So I will start with the lady at microphone three. Please proceed. I'm a lady, but that's OK. Yeah, it's OK. Thanks for your work, but I'm going to break the joy. No thanks for your talk. I find it in several ways racist and I want it to be said. OK. OK. All right, that's fine. Thank you to Frim. Yeah. OK. Please. So I would like to ask, given the work you have done, what is the legacy that has remained now that the Ebola epidemic is over? The systems that were built remain. So, excuse me. Yeah. So at this point, if I can just come in to answer the question. So what's happening now is the date of the disease has been transitioned up to the Ministers of Health. About 30 minutes or so before this, I was in the meeting with representatives from the Ministers and living representatives from e-Health since, you know, they are the ICT partners of the Ministers of Health. So the infrastructure is still very much in place there, finding new ways to, you know, to utilize it. OK. Great. Next? Next one. Number four, please. So this is kind of an evil question because after all, it was Ebola. Many people died. But do you think that maybe this crisis was, in some way, a good thing because we have now infrastructure and Sierra Leone and things are better than before? Well, in terms of things been better than before, I would not say it is. But at least now we understand what should be done. And I'm guessing, you know, the authorities understand how to respond better if there's something like this happen, things like this happen here on a smaller scale, you know, almost every week, there are immediately the breakouts, there are, you know, the descents, the breakouts and so on. And now I think, you know, the infrastructure to rapidly respond to these things is there. You know, from any bad thing, there are the good aspects, the good lessons that can be realized. I just hope it's been learned in our case. Thank you. Thank you. Anybody else? Yeah, person? Hi. I guess this is more of a business question. But with so many interested parties, you mentioned Nat Hope being a conglomeration of 30-plus NGOs, UNDP, IMF, World Bank, et cetera, et cetera. Having to move so quickly, what sort of reporting did you have to do? Did you have to get things cleared? You mentioned, you know, millions of dollars coming in every week. That's a lot of money, and people usually like to have some sort of standard on that. Yeah, so I can answer that. So there was a lot of reporting. And honestly, to get the systems developed at the right kind of speed and at the right kind of pace, what ended up happening was that many of the systems and checks and balances and controls were, they weren't bypassed. They ended up being abbreviated. Because, for example, any job that's over about $100,000, you're supposed to have competitive bids. If you're talking about trying to build an enterprise ERP in three weeks, there's simply no time to do that. So what ended up happening was that we told Sultan and the guys from my IT labs, you go ahead and start developing all of this stuff, because it has to be deployed right now. We will go and manipulate the system to enable that to happen. And in the meantime, the government, World Bank, IMF, and the rest of it will work through the finances to prime the pump so that when the system is ready to be deployed, the money is there. Everything collapsed together. I mean, this was a real emergency. The one other important thing that's probably worth saying is that there was no template for this sort of thing. Nobody knew how to do it. Nobody knew how to manage it. And in many cases, we had to invent everything on the spot. Massive credit is due to Sultan and his guys and the teams on the ground there, because they had to invent the business process. They had to invent the management on the spot. Even here, that would be an impossible challenge. Thank you. Thank you. And I don't really care if you all think I'm horribly racist or whatever, because I have a huge amount of respect for these guys. They save their nation. Signal of hope. Thank you so much. I think we can wrap it up for now. We'll run out of time for questions. Thank you. And if anyone wants to find me, I'll be in the smoking lounge. Yeah.