 Hiya, I'm Gavin Mendogleson, I'm the Chief Technology Officer of Terminus DB and I'm here to speak to you about CAMHS, which is the Critical Asset Management System. So CAMHS is designed for building climate resilience and creating a local knowledge graph for Dominica and beyond. So Dominica is a small Caribbean island nation and they are rich with natural resources and beauty, but they have quite a lot of problems with climate. So in 2017, Hurricane Maria, a Category 5 hurricane came in, 160 mile per hour winds, 90% of the island's structures were destroyed and 1.3 billion in losses, which is 224% of Dominica's annual GDP. So the Prime Minister, Roosevelt Scarrett said, the winds are merciless, we shall survive by the grace of God. So they really suffered under some extreme stresses. Addressing the United Nations, Prime Minister Scarrett appealed, I come to you straight from the front line of the war on climate change. In the past we would prepare for one heavy storm a year, now thousands of storms form on a breeze in the mid-Atlantic and line up to pound us with maximum force and fury. So world leaders came together to form the Paris Agreement in 2015, it's a UN Climate Change Conference COP 21 and the aim was to keep sea temperatures increases well below 2%, but do we trust world leaders to actually fulfill on those promises? And I think the reality that we'll find is that these boundaries will be crossed. So we're actually not doing anywhere near enough to stop climate change. So we have to be thinking about mitigation strategies as well because we can't move all of these parts at once. So the United Nations Office for Disaster Risk Reduction, the UNDRR, they've been looking at mitigation strategies and assisting countries with the strategies for that. So there's the Department of Planning and the Department of Industry and Commerce and Climate Execution Agency for Dominica. In conjunction with the RISE US, which is a private sector alliance that's working with the UNDRR on dealing with some of these climate mitigation strategies. So RISE is formed, it's a coalition of different actors. So Terminus DB is a partner, Data Sequia, BGC Engineering versus Labs, Tony Carato and Eric Su. So what are we building? We're building CAMHS, which is the critical asset management system for building climate resilience. And what is the goal of CAMHS? So CAMHS' goal is to get better visibility of what assets exist, what their dependencies are, and finding mitigation strategies so that when there are events that have occurred or are about to occur, there's some way to form planning decisions by having better visibility of what assets exist, what kind of status they have, and what the interrelationship of those are. So what does that include? That includes all key infrastructure systems, all emergency response assets, assets that can become critical in an emergency, for instance, schools are sometimes used for people to house them as temporary shelters, there's also backup power supplies, etc. There's assets or features that are a long way away, dams, levees, energy gridlinks, but that are related and important to understand. There's also natural systems, assets that are important to economic recovery, for instance, ports and other things. Assets are important to community recovery, such as community centers and welfare offices, and we need to be thinking about plans affecting all of these. So what's the problem to solve? At the moment, usually countries, utilities, all these different actors exist with these different assets, and some of them have partial understanding of what these assets are. So an electrical company might understand where all their pylons are, they might understand where their power sources are, but not everybody has visibility of these, and the impact of that on the grid is not always understood by everyone. So the way that these assets are also maintained is in a very wide range of different kinds of formats, so there's everything from asset management systems that are used primarily for understanding capital assets that exist inside of utilities or in corporations, if you have private corporations doing the phone infrastructure, for instance, as they do in Dominica, through to the state, which in Dominica will often have their assets listed in an Excel spreadsheet or in a CSV, or maintained even in books and things of that nature. So the problem that we need to do, that we have is actually we have to unify a lot of this knowledge from different sources with different requirements, such that you can get an overall knowledge graph that can give you a view of the information that you're interested in. So this gives us knowledge about the asset relationships together. So TerminusDB, that's where we come in, we're an open source, in memory, document oriented graph database, and we are designed for rapid application development for complex knowledge graphs. So when you write something with TerminusDB, first we step in by defining the sort of scope of the project, and we did this in conjunction with many different stakeholders in Dominica along with a rise. Then we build a schema, so the next stage is trying to figure out what kinds of assets you want to have visibility of, how those assets are connected to each other, what kind of additional information you need to store about those assets. And we do this collaboratively as a sort of hackathon session where we defined all of the elements that would go into this. Then we end up with a fully formed schema from this communication and this schema can then be browsed by non-technical users to understand the kinds of properties and connections between different elements. And here's just some views into our online user interface that you can browse the schema and see what kinds of elements exist for an asset. There's asset history, identifiers, commissioning date, description, design standards. There's a set of the last maintained time, so you'll have last modified and you'll have its location in space. As well as a spatial web identifier, which gives an absolute identifier for spatial coordinates in three dimensions. We can also see all the relationships that an asset holds with other things, so there's graded hazards, there's assets, there's events, so events might be something like a hurricane, graded hazards or information about what kinds of hazards can exist and what their gradations are. And this is just sort of an example of the forms that are generated automatically once you have that schema so that the information can be browsed from Terminus's dashboard. So the next stage is to build the UI, so once you have that, you automatically get these forms for entering in the documents, for entering in all the information and the connections that exist between documents, but it's not themed, so usually when you're building an application you want to theme it for your specific application, so we have a framework for building the UI where you just specify some UI JSON that UI JSON can even include callbacks to JavaScript applications that do more sophisticated changes, you can change the class structures in the CSS, etc., and allow it to be themed as you like. Then you can build a very simple and straightforward React application by referring to these types that you want to edit, that you want to update, or that you want to view. And then you can perform very sophisticated queries over the graph that involve lots of different data points linking them together, which is very difficult to do with just a document store like Mongo, but also quite hard to do with something like SQL, where relational databases make it somewhat difficult to do complicated chain queries. So we use data log query language, which allows you to concisely express complex patterns, these long chain patterns, and it supports both fluent and functional query styles. You can write things in a fluent style. It looks like, you know, it's a DSL essentially for JavaScript or for Python that allows you to do these graph queries in a quite native fashion. So the kinds of queries that we're doing, you can see the sort of graphy nature of it. Asset A depends on an asset B and it's dependent on an asset C. So like for instance, if you are a hospital, you're dependent on the power supply coming to you. If you're a pile, you know, if you're some intermediate producer, then you might also produce something for asset C that's required for things to continue. And each of these assets can be impacted by events. And they're located in a specific area and the events can take place in an area. And the events severity or scale is also important for whether or not the asset is going to be susceptible. So certain kinds of assets will be more or less accessible, or impacted by different scales of events. So an applicable, okay, so this is another good example of just how complex and graph-like the application actually is in the end. So you have assets, asset B can be dependent on, or asset A is dependent on asset B, which is dependent on asset C, which is dependent on asset D, which is dependent on asset E. And these dependency chains can mean that asset E could be affected by some kind of event that is not directly susceptible to, but that one of its other dependencies is susceptible to. So you have to know about the applicable hazards, the asset category, the asset owner's information about the maintenance history is also important. So you know how long ago it was maintained, and that might actually mean that it's not usable. This is really important, for instance, with backup power supplies. If they haven't been maintained, they may not actually be functional. And that happened in Puerto Rico in the floodings that happened a few years back. They actually thought they had backup power supplies that had not been checked. And so they didn't actually work when it came down to the problem. So this is a view on the actual CAMS application, which is, by the way, an open source application. Very easy to get started with. You can just pull it from our GitHub repository and then run it inside of a Docker. And then you can play around with it, modify it, and do what you like, open source. So this shows you some of the assets that are listed there. And when you clicked on this Trafalgar hydropower plant, you can see all of the things that are dependent on the Trafalgar hydropower plants. You get a list of the critical links on the side and all of the assets there. You can click through those assets and see their asset pages, all the details about it. You can also, you'll see at the top, there's a failure path and upward links. So you can see all of the things that it depends upon or all of the things that depend on it in a transitive closure. So it'll jump all of the hops necessary to see what those links are. And then you can navigate to these various different things. You can see where they're located, the information about them, or click through to them to get their information. You can also pick from applicable hazards at the top and see which hazards, at which category level are susceptible. So if there's a hurricane three, or a hurricane category three coming in, you could select hurricanes and typhoons, click three. And you could see all the assets that are liable to be severely damaged by a category three, but not those that would be susceptible to a category five. You see here that even in a category one, the airport ceases to function. And that could be interesting information from an asset view. So when you're loading assets, you can go, you can click in and see a sort of synopsis of the information about the asset. You can edit it. You can enter in new information, commissioning date, design standards, et cetera. And you can set up links for dependencies for that asset. So this is basically designed as a minimum viable product that would work for Dominica to deal with their problems. But we're also trying to expand out this use case and try to get other players involved. So there's a number of other new cities and countries, small countries, have been looking at using camps and developing it further. So what are the use cases for the minimum viable product that we see at the moment? Really, it's for planning during impending danger to really take an overview of asset vulnerabilities so you can make plans ahead of time by having that overview. And then for post disaster mitigation. So once things have actually gone down, it could be useful to know which of these assets is most critical to getting more assets up and functioning again. So what's next? Well, there's a lot more to come after this. So firstly, we want to deliver this MVP to more islands, cities, communities who need to be resilient to the climate change that is coming to them. And secondly, we want to have a lot of new features added that have been asked for by various cities and countries that are using it. And it is a collaborative open source project. So there's lots of interest from various different programmers, but we invite more people to come in and help us to actually make these features available. So one of them is automated alerts to asset owners. So sometimes the people running the camp system may not be the people who are actually the asset owners. For instance, in the case that you have an electrical provider, they might have information about the electrical power plant. And they can be warned by the people running the camp system who may be the state about the fact that they're susceptible to some kind of impending crisis. We'd like a mobile application built out of this. So a lot of the people who are going to be surveying or updating the assets or doing maintenance would like to be able to enter the information in from their phone or from a tablet. So we really think that it'd be important to have that and to make it so that it works offline with push notifications. So that all of the changes that they make would get pushed upstream once they actually are able to connect. Because many of these places may not have connectivity by cell. So that's also important. And lastly, we want more on analyzing the graph. So post disaster review screens, various different kinds of dashboard sorts of analysis that can be done. And to get some kind of ideas of changes over time and the impact of disasters. So that's CAMS, that's our open source project for good. And I hope people who are watching this who are interested in climate resilience come to the CAMS and try to start participating in the community. We have a website, a climate resilient.world, and it points through to the CAMS GitHub repo where you can get a hold of CAMS yourself so that you can contribute to it or use it if you're looking for a asset management system for a small city or even for companies that might want to do asset management in this way, they can have a look at that. So this is our GitHub critical assets management group. And we also have a board that has all of the kinds of issues where it is in the process of the pipeline for in progress done, doing, and who's actually doing that. So thank you for listening and do try to get involved if you can. Thank you.