WEBVTT 00:00.000 --> 00:19.760 Hi everyone, thank you for joining us for this talk today, I'm very excited to demo our 00:19.760 --> 00:28.880 prototype climate health pulse. My name is Tali and I am from Data Kind. I have a few other 00:28.880 --> 00:37.760 colleagues here as well from Data Kind. We are a global nonprofit organization and we have 00:37.760 --> 00:45.520 a reason for existence, if you like, is mainly to close the technology gap that a lot of social 00:45.520 --> 00:52.320 impact organizations face when working with data science and working with AI and essentially 00:52.320 --> 01:00.400 we, the way Data Kind works is we collaborate with amazing experts like yourself and we work 01:00.400 --> 01:06.960 with highly impactful organizations in particular sectors to then design and develop products 01:06.960 --> 01:12.720 that are sector facing. So it's not really focused on just solving bespoke problems for a partner 01:12.720 --> 01:17.840 or social sector organization, but really trying to understand the pain point that impact the 01:17.840 --> 01:23.760 sector as a whole and then work with a few key actors to actually design and develop those products 01:23.760 --> 01:32.160 that are then available as open source products to the sector as a whole and it is it is our 01:32.160 --> 01:38.160 mutual commitment to building a more just world through responsible and ethical use of data science 01:38.160 --> 01:47.120 and AI. So the solution that I am going to demo today has been developed in collaboration with 01:47.120 --> 01:52.320 such partners. So I want to just name them, they are not here by I want to highlight Spectrum 01:52.320 --> 02:00.080 Africa. This is a Canyon-based organization that they work to bring together multiple sectors 02:00.640 --> 02:04.960 to work together on uncovering sort of the hidden connections between climate and health. 02:05.360 --> 02:13.200 And climate health and also generally well-being of people, which as we know has more implications 02:13.200 --> 02:20.320 than just health. I also want to mention Jackarander Health. They partner with governments, 02:20.320 --> 02:28.960 public hospitals, and with pregnant and expect, well, pregnant and new mum directly to design and 02:29.040 --> 02:36.480 develop products to support their mum's on their journey on the pregnancy journey and through 02:36.480 --> 02:43.920 motherhood, the early stages of motherhood. Their flagship platform is prompts and prompts 02:43.920 --> 02:51.280 supports over nearly actually nearly 3 million subscribers at this point, new and expecting mothers 02:51.280 --> 02:58.800 through a two-way SMS messaging platform and they basically send stage specific messages to 02:58.960 --> 03:05.920 those moms and they also respond to queries from those moms. And we wouldn't be able to show you 03:05.920 --> 03:14.400 this demo today without the cooperation, without the support of Cardiota County. This is a 03:14.400 --> 03:20.480 county in Southern Kenya, the Ministry of Health specifically from this county, has partnered with 03:20.480 --> 03:25.920 us to be able to share their data and show you what's actually possible to do now. 03:26.880 --> 03:34.800 Just a little bit about Cardiota County, it is beautiful, it has valleys and plains and it's 03:34.800 --> 03:45.200 sort of a narrowed kind of climate, home to 1.1 million people, mostly for the majority, a 03:45.200 --> 03:52.240 pastoralist community. You might have heard of the Masae, Mara, the Masae, they predominantly 03:53.200 --> 04:00.960 live and work in this area. And they face their share of climate challenges, they really struggle 04:00.960 --> 04:10.080 with heat, stress as well as water scarcity. All right, so climate change in Cardiota County and around 04:10.080 --> 04:15.760 the world is really, it's a current and escalating health crisis. We have recurrent droughts, 04:15.760 --> 04:22.880 we have their ethnic rainfall, heat waves, they directly impact health. And in some cases, climate 04:22.880 --> 04:29.120 changes sort of impacting the, it's increasing the risk of certain climate sensitive diseases. In 04:29.120 --> 04:34.800 other cases, it's changing the way those diseases manifest, it's making it, the seasonality of 04:34.800 --> 04:41.360 those diseases, it's just making it harder to anticipate. When certain diseases will strike and 04:41.360 --> 04:46.560 what to, and to what extent. This is known, we know it, health systems overall around the world, 04:46.560 --> 04:53.840 they know it. And at the most local levels, they're facing health systems are really facing the 04:53.840 --> 05:01.760 strain of these health crises. So it's no longer a question about how we use climate data, 05:01.760 --> 05:06.480 how do we better use publicly of it, you know, just how do we use climate data for health, but 05:08.240 --> 05:15.120 what can you do with that, what kind of action could it support? I just want to mention that 05:15.920 --> 05:22.080 the way we build our products, our data kind, it is based in understanding the needs of the 05:22.080 --> 05:27.040 sector, I mentioned that earlier. So last year, data kind spent about 10 months in dialogue with a 05:27.040 --> 05:33.040 variety of different stakeholders, including data providers, including developers, including 05:34.640 --> 05:39.840 folks who are actually in the decision-making space when it comes to health systems, researchers, 05:39.840 --> 05:44.560 policy makers, and so on. And also actually frontline health organizations like Jaco and 05:44.560 --> 05:51.120 the Health, who partner with governments to assist in strengthening those health systems. So 05:51.120 --> 05:56.800 we spoke to them there from across African Asia and Europe mostly, and some in North America. 05:56.880 --> 06:02.720 And what we found was that health stakeholders, especially across the, you know, thus global 06:02.720 --> 06:10.480 south, definitely recognize the need for an actively seeking climate informed health data 06:10.480 --> 06:16.880 to support the decision-making processes, but despite that demand, there is a critical disconnect. 06:17.600 --> 06:24.800 I think this was actually mentioned in a few presentations, but the data is actually quite segmented 06:24.800 --> 06:29.120 and the formats are such that it makes it quite challenging for folks in the health sector 06:29.120 --> 06:38.800 to use data from let's say the climate sector. Also, the climate health, yes, have generally 06:38.800 --> 06:44.240 traditionally worked sort of in silos. So bringing them together is one part of the problem, but then 06:44.240 --> 06:51.840 even when that data is available, that localized near real-time climate information is rarely actually 06:51.920 --> 06:57.200 integrated into existing workloads. And that's where it could actually have the most impact. 06:57.760 --> 07:03.680 So for what we found was that this really represented a significant missed opportunity to strengthen 07:04.240 --> 07:08.000 health system resilience through better integration of the data. 07:10.240 --> 07:14.880 You can actually, if you're interested, you can read more about that report and the challenges 07:14.880 --> 07:20.880 that limit climate and health data integration in our landscape report, as well as 07:20.880 --> 07:25.440 there, we've also noted some promising innovations from around the world. In our report, 07:25.440 --> 07:30.960 you can use that QR code to get there. And what you'll note in the report is while the ecosystem 07:30.960 --> 07:38.160 is fragmented and siloed, government agencies are absolutely central. They are absolutely central 07:38.160 --> 07:44.480 to enabling that data access and the data utilization. Questions around data availability, 07:44.480 --> 07:49.840 interoperability and capacity, they are persistent, yes, but actually solvable. 07:50.720 --> 07:55.280 And while resources are limited, we can still focus on feasible, cost-effective 07:56.240 --> 08:03.520 open-source approaches to strengthening existing capacity. And this is coming at it from an equity 08:03.520 --> 08:09.760 lens as well. So we wanted to find out then, okay, what if we were able to actually integrate 08:09.760 --> 08:16.320 climate and health data? Rather than developing some kind of a narrow tool, we are working on 08:16.320 --> 08:23.440 an integrated platform that can potentially unlock that more ecosystem's level thinking. So 08:23.440 --> 08:27.840 enabling more sectors to come together, more sectors to collaborate, bring their data onto 08:27.840 --> 08:36.480 onto a shared platform and enable iterative learning. So that is our first step towards that is 08:36.480 --> 08:46.240 climate health pulse. It, in its current form, it brings together climate data and health indicators 08:46.240 --> 08:53.840 at a granular county level. And it actually levels below that. So in Kenya, you have the 08:53.840 --> 08:59.920 national, you have county sub-county and ward level. So this is looking at county and below. So 08:59.920 --> 09:04.960 more, as localised, we can possibly get when it comes to administrative boundaries. 09:06.960 --> 09:13.280 It exists in this iteration again to make climate and health data useful, usable, actionable 09:13.280 --> 09:19.600 for local governments. And really understanding risk is one part of it, yes, but also opening up 09:19.600 --> 09:26.640 those opportunities to act on it. What we really wanted to show is that pulse is not just about 09:26.640 --> 09:31.920 like a data source, so bringing together some, you know, a couple of layers, but it's really 09:31.920 --> 09:38.320 about layering on multiple sources of information from multiple sectors to create that actionable 09:38.320 --> 09:49.360 climate related intelligence on which you can work. Pulse is a solution that is designed to 09:49.360 --> 09:56.240 evolve with new partnerships. This is, this is by design. We're working on it with that. It is 09:57.120 --> 10:01.760 we need new partnerships. We need new data. We need new indicators, new geographies. And I 10:01.760 --> 10:06.640 think we are here for our data kind and part to learn about other tools as well that we should be 10:07.360 --> 10:12.640 aware of that we should be using as well. It's being developed as a free and open source 10:13.520 --> 10:21.200 platform. A data kind we do see open development as the most effective way to maximize transparency 10:21.360 --> 10:28.160 reuse and adaptability. So really ensuring that the tools we build can be sustained, improved, 10:28.160 --> 10:34.720 and deployed long after sort of individual partnerships fade away and, you know, and that 10:34.720 --> 10:40.800 ends. We're also developing Pulse to digital public good standards. And there is, I think, 10:40.800 --> 10:48.160 if, I don't know, if anyone hears from the DPGA, right? The digital public goods alliance, 10:48.160 --> 10:52.960 they have a stand outside as well. It's a really important, particularly when we're working with 10:52.960 --> 10:59.200 government actors to enable them to use open source software. It is an important sort of accreditation. 10:59.200 --> 11:05.040 There is a trust involved in using digital public goods that are open source tools and platforms. 11:06.000 --> 11:10.640 So we use those standards. We devote to those standards with the goal that even if it isn't 11:10.640 --> 11:13.200 certified right now, it's on the path to that certification. 11:14.160 --> 11:23.040 Okay, with that, I'm going to go into the demo. Let's see, I'm going to do this a bit backwards. 11:24.960 --> 11:40.000 Right. So we can just get started. This is the main page. It just contains, you know, the 11:40.000 --> 11:48.320 information that I've already shared with you just now about the project. As we scroll down on 11:48.320 --> 11:55.840 right here, you get to our featured prototype. And basically, this does actually, because we're 11:55.840 --> 12:01.280 talking about Cardiata County, just wanted to highlight, again, the different data sources, 12:01.280 --> 12:07.040 so in front of managed two screens here. Okay, we have health data from the Kenyan Ministry of Health. 12:07.040 --> 12:12.080 This is private data. So it's not right now. If you try to go on there, you can't actually access this. 12:13.120 --> 12:18.240 This tool we are working on a open data set that's going to be available and then we'll be able to open source 12:20.320 --> 12:27.360 project here. We also have not just sort of health, you know, from the health informatic system. 12:27.360 --> 12:31.360 We have the data, but we also have facility list. So really getting down granular to the actual 12:31.360 --> 12:37.600 health facility, whether it's a clinic or hospital in Cardiata County. There are 116 active 12:37.600 --> 12:44.000 government facilities in that county, so we have that level of data as well. All anonymized, we do 12:44.000 --> 12:49.440 not look at the actual patient records right now. We have climate data source from Copernicus, so 12:49.440 --> 12:54.640 for the last group that was here talking about that. It's the error five data set, which is a 12:54.640 --> 12:59.360 really crucial data set for climate research, offering consistent data on temperature, 13:00.080 --> 13:06.320 precipitation, wind, and more at a fairly good resolution of 9 by 9 square kilometers. 13:07.120 --> 13:13.200 And we also have actually four cast data coming in from the Kenyan meteorological department 13:13.200 --> 13:19.120 directly. So we're actually working now closely to build that local relationship with K-met 13:19.120 --> 13:23.520 and make sure that they are the ones who are also providing the climate data. 13:24.560 --> 13:28.480 And then finally we also have the Jack Render Health prompts data that I mentioned earlier. 13:29.360 --> 13:34.880 Mums from across the country are asking thousands of questions about their pregnancy and 13:34.880 --> 13:40.240 about their babies. It's a really valuable data set. They've shared some subject matter 13:40.240 --> 13:46.080 classifications of those text messages aggregated at world levels and different levels. 13:47.040 --> 13:52.240 But this data gets us a community report. It's really the first time that government, 13:52.240 --> 13:57.360 we're working on a tool that not only has a government data and other kind of globally available, 13:57.360 --> 14:02.720 like open data sets, but really trying to bring in community data as well. So how people actually 14:02.720 --> 14:11.520 experience climate. So if you go into the actual tool, 14:15.440 --> 14:22.080 all right. So this is the overview page here. We have essentially four layers of information 14:22.720 --> 14:31.840 that you can see. It's not showing the rest anyway. It looks a bit different on my computer 14:31.840 --> 14:37.200 than there. So this essentially shows Kajada County's map. You see the different world levels. 14:38.800 --> 14:47.200 The color to represent sort of the disease incidence rates and so on is just a fairly straightforward 14:47.200 --> 14:53.840 and easy way to visualize patterns. Identifying hotspots, that kind of thing. It's super helpful. 14:53.840 --> 15:00.160 On the map, you can kind of see, I think, the health facilities we have like the actual, 15:01.120 --> 15:05.200 so you can actually see we've got the health facility locations mapped on as well. Right now, 15:05.200 --> 15:10.000 this just has very basic information about that health facility. But for the actual government user, 15:10.320 --> 15:17.280 the end view that they would have detailed information about that health facility, 15:17.280 --> 15:23.680 including specific for specific indicators, what are the rates and so on. 15:24.720 --> 15:30.400 Okay, and then next down below here is a simple line chart. Again, just showing that health outcome trend 15:30.400 --> 15:35.600 for any given indicator. Over time, this is also extracted from the Kenyan Ministry of Health 15:35.760 --> 15:42.800 Information System. On the here, we have climate information, so total precipitation, 15:42.800 --> 15:48.400 average daily max temperature. This is that Copernicus data set and then below that a simple 15:48.400 --> 15:56.240 line chart again showing those community reports, the prompt data and you can actually like search 15:56.240 --> 16:02.720 those as well and kind of look at the trends there. Some additional features involve, you know, 16:02.800 --> 16:07.360 you can actually select an indicator. So again, you know, you can actually pick your indicator. 16:07.360 --> 16:10.640 I'm going to start looking at malaria in a second, which we know has quite a bit of 16:12.320 --> 16:17.120 relationship with climate. So there, you can look at it for 16:20.720 --> 16:22.640 different time periods. 16:26.560 --> 16:29.520 Really see what's happening. This is now over almost three-year period. 16:29.520 --> 16:35.520 And depending on what you find, it's got a few other key features just to help. You can download 16:35.520 --> 16:43.040 these images in different formats. And yeah, so you can look at it and you can actually drill 16:43.040 --> 16:48.000 down further. You can actually go from that county level to different sub-county level or 16:48.000 --> 16:56.960 word-level even and look at the data more locally. So if you look at malaria, you know, you 16:56.960 --> 17:03.680 set that data in. You can actually have a sense of what's happening over time when you play 17:03.680 --> 17:09.520 it out. And you'll just see that there's a the north part of the country. This is just based on 17:09.520 --> 17:13.200 this is a normalized data set from what we received. You can see the north part of the county 17:13.200 --> 17:20.400 really in general has particularly high confirmed cases. It's a hot spot for malaria. 17:21.280 --> 17:26.000 You can get closer into that and drill down and see what's going on particularly in 17:26.320 --> 17:32.640 this what here you can zoom in and have a look. And the end user would be able to actually see 17:32.640 --> 17:37.600 what's going on at the specific health facilities. The more important thing is though on the right 17:37.600 --> 17:42.880 you actually get to see what's going on and what's been going on with climate, what's been going 17:42.880 --> 17:48.400 with temperature, what's been going with rainfall over the same period of time. And you'll notice 17:48.400 --> 17:53.120 that all the other layers are just, you know, as you sort of drill down all the other layers 17:53.120 --> 17:57.840 also adjust to give you the data about that specific what an area. 17:59.680 --> 18:03.280 Trendline really helps to see and understand like okay what's going on this looks like a 18:03.280 --> 18:08.080 particular, particularly high peak. You kind of see where the where the trendline starts to 18:09.440 --> 18:14.560 ensure an increase and you can actually start to just even that even just trying to look at 18:14.560 --> 18:20.640 what's happening with temperature and rainfall when there is a known where the trendline from 18:20.640 --> 18:26.080 malaria cases increasing. This is something that the local government acts is just weren't able to 18:26.080 --> 18:30.320 do. It was just not something that they could look at on just one simple dashboard. 18:33.920 --> 18:37.440 Further down you can actually have some use cases where we worked with the Cajillo 18:37.440 --> 18:42.160 to county health officials to look at specifically when there was a drought, what happened to 18:42.160 --> 18:46.960 malnutrition cases during that drought, what happened to rainfall and start to actually study 18:46.960 --> 18:52.000 and see what are the types of issues that months were raising during those specific events 18:52.000 --> 19:00.640 where climate events. I'll just show you a little bit speaking of that the prompts data here. 19:01.760 --> 19:04.720 Oh, look went to the wrong one. That one. 19:08.480 --> 19:16.560 So this enables users to also explore the issues cropping up on prompts at the time. 19:17.440 --> 19:22.480 It's an opportunity as an institution to actually learn more about community issues during climate events. 19:23.760 --> 19:28.240 There's an amazing study that Jackaranda Health have done that actually that's published 19:28.240 --> 19:33.600 that shows that prompts data tells us about how those millions of moms are experiencing 19:33.600 --> 19:39.680 extreme heat and we wanted to include that here as part of the perspective to inform the public health 19:39.840 --> 19:46.880 actions. So in this first set of visuals, basically we can investigate the relative distribution 19:46.880 --> 19:53.200 of those intense in sort of different temperature buckets and bins and we can see the changes 19:53.840 --> 19:59.840 in frequency of those specific intents. In general, I would say the questions about 20:01.040 --> 20:03.840 what we found questions about in studying this further. 20:04.800 --> 20:10.880 Pregnancy, newborns, questions about intercourse, questions about pains, swelling, urinary tract 20:10.880 --> 20:19.760 infections. Those kind of questions actually we see there isn't increase at higher temperatures 20:19.760 --> 20:24.800 in those buckets and the effect is modest but it is statistically significant. 20:26.640 --> 20:32.960 A couple more examples again when you search for different intents that we know are related to heat 20:33.040 --> 20:40.880 events or we suspect. There is a relationship here, for example, moms talking about swelling 20:40.880 --> 20:49.760 or a demo. When temperature climbs above 30 degrees, we see nearly 70% rise in those types of concerns. 20:50.800 --> 20:56.000 We see a 23% rise when we look at questions about labor and in other kinds of 20:56.000 --> 21:00.080 different health indicators for instance with children and babies, we also saw similar kind of 21:00.160 --> 21:07.280 relationships in the frequency of intense around things like diarrheal diseases, fever spikes, 21:08.080 --> 21:14.240 inability to feed and so on. What we found basically what we demonstrate to 21:14.240 --> 21:18.080 and this is again think about the institutional user in mind what we are trying to show is heat 21:18.080 --> 21:23.520 isn't just uncomfortable for a pregnant woman, it's changing what they need from the health system. 21:23.520 --> 21:27.600 So this is an critical input that wasn't actually available to 21:28.400 --> 21:37.120 those health system managers until now. I think this is also consistent when you have extreme heat, 21:37.120 --> 21:41.920 when you have a climate event, engagement with that platform increases. It's also part of 21:41.920 --> 21:47.360 the contributing factor as well to the kind of things we're seeing. But beyond just looking at 21:47.360 --> 21:56.880 sort of existing historical data and what's happened, the county health offices 21:57.920 --> 22:03.360 want to use data like this to think about the future and think about what can be done to 22:03.360 --> 22:09.040 adapt to build resilience, what strategies are open to them and this is where we were actually inspired 22:09.040 --> 22:16.880 by the work of Carriette Al, I do have the links available as well. We use generalised additive models 22:17.840 --> 22:23.040 to understand the relationship between temperature and precipitation and maternal and newborn 22:23.040 --> 22:28.720 health indicators. This was actually developed, this methodology was developed in the DRC, 22:28.720 --> 22:35.120 it's open source, it's available and we we use that in the Canyon context. To forecast 22:36.800 --> 22:42.560 the relationship between temperature and precipitation from Copernicus and the health indicators 22:42.560 --> 22:51.680 mentioned here. You can see there's a predictive forecast space that's really right now has 22:51.680 --> 22:56.320 a lot of uncertainty associated with it, it's currently not necessarily actually useful in its 22:56.320 --> 23:02.560 current format, but this is another area where I think we're very keen and open to have partnerships 23:02.640 --> 23:13.680 to bring more relevant data, that more relevant models validate those models and actually make 23:13.680 --> 23:18.560 this, put this in front of decision makers to think about really what can be done, can we think 23:18.640 --> 23:34.640 about more disaggregated ways of supporting health systems? So yeah, like I said we are looking, 23:34.640 --> 23:41.120 this is all about partnerships and cooperation community, we develop pulse and we're starting 23:41.120 --> 23:47.200 to develop those partnerships and community around climate health data integration, modeling, 23:47.280 --> 23:54.080 validation, please consider joining our community, you can select, you know, we have some options 23:54.080 --> 23:58.320 select ways in which you'd like to contribute, or suggest ways in which we haven't even thought of. 23:58.320 --> 24:04.320 I mean, we definitely need a lot of different expertise here to actually build this into a 24:05.200 --> 24:09.920 project that not only serves Cajillo to county and all the counties of Kenya, but actually 24:09.920 --> 24:16.720 make this into a project that can be then adapted to multiple different locations, again, 24:16.720 --> 24:22.880 really localizing the way in which we look at climate data and other sector-based data like health. 24:24.480 --> 24:29.360 So with that, I'll say you can check out the website, but sorry you can't see the tool just yet. 24:31.360 --> 24:32.560 Yeah, and thank you. 24:40.400 --> 24:44.880 Any questions? 24:58.320 --> 25:02.560 Do you provide public data sets for download? 25:03.360 --> 25:09.360 We're working on it, but actually we are for demonstrative purposes, yes, that's what we're 25:09.360 --> 25:16.640 working on, public like open data sets just to kind of see what the tool can do, but I think it would 25:16.640 --> 25:23.520 be quite interesting for us to actually provide public data sets that can be used in specific locations, 25:23.520 --> 25:28.960 or that can be tailored to certain locations. So yeah, we'd be open. Thanks. 25:45.600 --> 25:49.440 Hey, thank you for the valuable presentation. 25:49.440 --> 25:59.360 This career is dependent on you. What is the decision that this either should take in that 25:59.360 --> 26:03.360 particular case? What are the conclusions we can withdraw from this data? 26:05.280 --> 26:10.800 What brought us to build this in the first place? Is that your question? Sorry. Yes, the 26:11.760 --> 26:17.760 Yes, and this with the data we have now, what can we do with it concretely? 26:18.480 --> 26:26.800 That's a great question. So part of why this came about in the first place is because of who 26:26.800 --> 26:33.040 we are as an organization. So we have been working to support frontline health systems with open 26:33.040 --> 26:40.400 source tooling for the last six years, I would say, and one of the major issues actually that came 26:40.400 --> 26:46.320 up early on was around, we have a lot of data, we have digital data, but we don't trust that we don't 26:46.320 --> 26:51.920 want to use it. So data quality was usually what we'd come up against before we heard any other 26:51.920 --> 26:56.560 issues. So we initially started working on tooling, which is actually a digital public good, 26:56.560 --> 27:02.400 it's open source, it's called the data observation toolkit, or dot, and that is available right now 27:02.400 --> 27:07.840 for energy and use by these health systems. But as we got more interactive with the health managers, 27:07.840 --> 27:15.200 with the governments, more explicit use cases started to emerge, and this issue around climate 27:15.200 --> 27:19.040 is always there. Some governments are further along in terms of having a climate agenda. 27:19.840 --> 27:27.440 Kenya was particularly receptive to have a strong policy, but also want to show tooling that 27:27.440 --> 27:32.560 kind of interacts and informs that policy as well. So we found this really favorable environment 27:32.560 --> 27:39.040 for working with government partners on this. Plus, it also provided us, we've been working for 27:39.040 --> 27:43.760 several years and actually open source, well, it's a methodology that is now open source and 27:43.760 --> 27:48.160 available, but how to actually pair climate and health data. So we've been working with the 27:49.520 --> 27:55.760 UK's Office of National Statistics, along with Ghana, Rwanda, and a couple of other places 27:55.760 --> 28:00.640 on that methodology. So this was an opportunity now to start to bring all of those pieces together 28:01.360 --> 28:07.680 on a platform that could work on this. And then I think the second part of your question was around, 28:08.800 --> 28:13.760 sorry, you're going to have to remind me second part to say it all out. What they can do with it, 28:13.760 --> 28:19.840 thank you. So our partners in Kajeta County and from other counties who have so far seen this, 28:20.000 --> 28:27.520 the kind of actions they're keen to take is, for instance, being able to understand a little bit more 28:27.520 --> 28:32.160 of the specific variants of the diseases and the relationship to the climate to think about, okay, 28:32.160 --> 28:38.880 where can you send certain drugs in advance? Where can you impact supply chain, for instance? 28:39.440 --> 28:44.560 What can you do to support the community health workers when you know there is a spike coming up 28:44.560 --> 28:50.400 of certain types of diseases? That's one thing. You can also look at the community reports to 28:50.400 --> 28:56.320 understand the kind of questions that people, the community itself has. And so start to engage 28:56.320 --> 29:01.600 your community health promotion campaign messaging and so on, based specifically on the data 29:01.600 --> 29:05.680 and information getting from people themselves. So there are quite a few operational things that 29:05.680 --> 29:11.280 can be done. It's a matter of now seeing if they have the confidence in the data to be able to take 29:11.280 --> 29:21.200 that forward. Hi, yes. On that slide, on the ways that contribute, you mentioned red team 29:21.200 --> 29:27.840 experience. What does that look like in near use cases? I'm going to ask our VP of technology, Larry, 29:27.840 --> 29:44.640 to speak to that. As you can imagine with a prototype like this, before we publicly start 29:44.640 --> 29:49.440 putting it out for use and having people contribute, we are wanting to think of what are we not 29:49.440 --> 29:55.680 thinking about before we do that. And so when we say red team expertise, we're thinking about 29:55.760 --> 30:01.600 some of the skills that are listed here like data security experts in Kenyan health data and the 30:01.600 --> 30:06.000 rules protecting that. But also saying here are the things that you haven't thought about yet 30:06.960 --> 30:11.680 so that our implementation plan to go ahead and make this not a prototype, but the real tool that 30:11.680 --> 30:17.600 could be used, not just at the county level, but in all of Kenyan first and potentially in the region. 30:19.600 --> 30:24.880 We don't have necessarily all the expertise because of our individual focuses 30:25.920 --> 30:31.120 so what we want is for people who've done implementation, implementations like this to say here 30:31.120 --> 30:36.240 are the things that you aren't thinking about and then help us test it to make sure we've implemented 30:36.240 --> 30:42.080 all those things we haven't thought about and the ones that we've thought about. Does that? 30:42.400 --> 31:01.840 I have a question on the technical side. My understanding is that this tool is open source. 31:02.640 --> 31:07.360 Could you share a little about the text that used to create this tool? 31:07.360 --> 31:24.400 Sure. So currently, as you can see, we, all the way to the left, we are using some of the 31:24.400 --> 31:29.200 public data sources but we are still using some of the private data sources and working on how 31:29.200 --> 31:34.240 to make those available. But the actual tech stack is as you can see we do a bunch of processing 31:34.560 --> 31:40.160 and those are done with both Python and R scripts, shell scripts running the modeling. 31:41.440 --> 31:49.280 And then that is put into the data layer which is a post-gray SQL as well as a GIS extension since 31:49.280 --> 31:56.560 this is a GIS room. And then we're using on that back end layer, Net's JS, the Net's JS API. 31:57.520 --> 32:04.000 The front end is really pretty simple. It's basically react primarily and then for the maps here 32:04.000 --> 32:11.920 we're using leaflet and then all of that we're running right now on a GCP instance which is all 32:11.920 --> 32:19.440 pretty much a straightforward cloud run deployment connecting to a cloud SQL and then for the monitoring 32:20.400 --> 32:28.480 we're using a bunch of both native tools and some third-party tools to do that monitoring 32:28.480 --> 32:35.840 things like data dog which are commercial but to have open source APIs that you can go ahead and 32:35.840 --> 32:40.720 install and then of course the native stuff in GCP. So the primary parts of the stack are 32:40.720 --> 32:47.920 pretty much Python and react. Does that answer? As we go forward we'll be looking for more opportunities 32:47.920 --> 32:53.680 as Tyler said as we progress more on the modeling is how can we integrate more modeling options 32:53.680 --> 32:58.400 in a better way. So let's start to look at some of the options that come with some of the 32:58.400 --> 33:01.760 different modeling packages that can do real-time selection of modeling. 33:04.800 --> 33:06.560 I think we don't time for one more quick question. 33:06.640 --> 33:19.200 Thank you for the for the nice talk. I just have three quick points. One is there was recently 33:19.200 --> 33:24.960 in the news the US bought all the canyon health data from the government. Do you have any 33:24.960 --> 33:30.480 comments on this? How does the fact what you're doing? And two is that I am a member of the 33:30.480 --> 33:35.520 project called HealthSight.io and maybe we can talk afterwards if you would like to collaborate. 33:35.520 --> 33:41.520 Health sites? Health sites? Yes. And three I'm a member of the project working on something 33:41.520 --> 33:46.400 called tomorrow now where they're doing climate forecasting and maybe it's also something we can 33:46.400 --> 33:50.720 chat about. That'd be amazing I would appreciate. So just the one question was about the 33:50.720 --> 33:57.520 US buying all the health data? Yeah I don't think I'm nothing specific to comment right now I 33:57.520 --> 34:04.960 think we are also trying to understand the implications. Any more questions? I think we're still 34:04.960 --> 34:16.320 about time for one quick question if anyone. Okay. Yeah. Thank you very much.