 All right. Hello, everybody. As I said earlier, today's going to be all about you. I don't have any additional slides other than what we saw this morning. We can ask questions, but I also would like to kind of go around the room here and have each country or each group that is here talk a little bit about how you're using maps, how you would like to use maps, what is maybe preventing you from using them the way that you would like, or just ask questions. So we'll be going around the room. So start thinking now about what you want to talk about or what you want to say. And then we'll have a little discussion with each group. So it'll be more of a discussion session and more of a working session. If you have questions or if you want to try something together, we can do it on the screen as well. If you want to test out some of the functionality that I showed this morning, if you have questions about anything or you want me to go a little bit deeper into what it actually means, we can do that together. Yeah, that's how we're going to start. That's how we're going to go about this. But first, I want to go to the slides that I had at the end of the last session. So in this case, we're looking at maps that were created by some of the participants in the Maps Academy in South Africa earlier this year. And I want to go together and hopefully we have a mic that we can go around with so that we can try to interpret these maps together. So I'm not going to tell you what these mean. There's some text that kind of explains a little bit of it. But hopefully the map itself tells the story. And this is really important when you're using maps or designing any sort of data visualization is that it should tell a story. And you can tell different stories depending on how you configure, how you set up your visualization or your map. But if you do it well, it should tell you what you're trying to learn from this map. So let's start with this one. And I know that we're a little shy on volunteers sometimes when I asked for a raise of hands earlier. But don't be shy. Please, does anyone want to take a first attempt and we can work on it together to interpret? What is this saying? So what kind of, what story are we being told by this map here? Who wants to start? OK, if we look at the legends, ANC coverage, and actually the intensity of color shows like how many coverage like we have in particular area. So I don't see the labels of the area. So and this map is even not my country map. So this is only what I get is like it's showing the coverage. Yeah. So that's that's good. That's that's actually there are two layers here. If you see there are actually two different three, but there are two different legends down there. And and one of them maybe it's hard to hard to read. It's actually written up here, though. So we have two different layers. One of them, as you mentioned, is ANC coverage. And the other one is women of childbearing age. So number of women of childbearing age. So those are related in some way, right? And so as you can see, it's actually described here as well. And the map is used to identify where the women of childbearing age are located. And then we can design interventions to target them using ANC programs, for example. And again, these were created by people from these countries. So this is their take on how they would like to use maps and applications. There might be other ways to do this. This might not be the way that it would work in your country. But it can tell you can tell a little bit of a story. So you can see where kind of the the targeted outreach for ANC coverage improvement would need to be for this country. Thank you. OK, let's move on to Lesotho. And we actually have an explanation that we can go to in a minute. What is what story does this map tell you? And I'm also interpreting these with you. So we'll try to hope we can get the story that was being told by the map authors, which was not me. I think the number of home deliveries, actually, this is the bubble map. And it shows the number of deliveries, the larger the bubble and the intensity of color is indicating that the number of deliveries is more in the larger bubble. Yeah. So there's a lot of home deliveries in this area right here, but fewer in the other regions. Let's see what their explanation said. And let's see if other people can help contribute to this as well. Oh, this is just a picture. Hopefully. Yeah, there we go. Sorry. Yeah, do you have a? Yeah. Okay, I don't know if you guys could hear that super well, but it's really interesting what you can do when you try to look at a map and understand what the data is actually telling you, what it actually means, right? So we could see in this example that there's a lot of home births in this hospital. But what they identified is if you look at the elevation, you could even use an elevation layer to see that there's mountains and very, very rough terrain nearby and very rural areas. And that is actually where the home births are happening. So people maybe don't have access to a health facility or the hospital where they're doing most of their in staff deliveries. And so they're doing, they have many more cases of home births in those cases. And so this might be a way to do outreach to those communities to work on access or work on facilities that are available to communities in that area. So there's a lot of interesting things that you can take, not just first seeing, yes, there are a bunch of home births there, but what does that actually mean and why is that happening? Okay, before we go on to looking at some more of these, I do want to do the exercise of going around and every group, talking a little bit about how you're using maps today, maybe how you would like to use maps and maybe any of the features you saw today or features that DHS who doesn't support that you think would make it be helpful in the programs that you're supporting. So maybe we can go around the room and start one table at a time and do that. Maybe start on this side since you were such a brave volunteer earlier. Thank you. Oh, sure, yeah. I was giving you a break and going around the other way, but that's all right. Thank you. Should I explain what we are doing in our country? Yeah, if you want to introduce yourself quickly. Yeah, actually, me and Janze, like my fellow, we are leading IDSR implementation in Pakistan and also presenting National Institute of Health here. We are using maps and we imported the shapefiles into DHS too and right now we have a province level, the district level, the sea level and UC level boundaries available in DHS too. And we are also providing training to every province to have their own dashboards. And in the dashboards, like they are using maps to indicate diseases and the intensity of disease in particular area. So that feature is well and always they ask us like, they want to click on the boundaries and they want to drill down within the maps and that feature is not available in DHS too right now. Click on the boundary and go down in the maps. Actually they want like, you can say if I'm watching a district and the district is showing the number of diseases. No, I want to click on the map so that it can bring all the tassels and the uses and the health facilities. Yeah, maybe they want to do a drill down function within the, yeah, that's how this is missing. Yeah. And the other thing which I saw right now we can present two diseases at a time on the map as well. Go ahead. No, yeah. So does anyone have comments on that or suggestion? I think that feature might actually already exist. So I believe that we do have drill up and drill down functionality within the maps application. So we'd have to dive in a little bit more to see if it matches your use case but I'll just show it quickly since we have it here. The internet is a little bit slow, apologies. Maybe we are using 2.36. 2.36, it should be there in 3.6 as well. Yeah, let's go. So we can, I don't think we have 3.6 on play but I can, I can spin up. In events we can do, I think they're talking about the thematic layers if they can click and drill down level below. Yeah. Yeah, in thematic layers you can. So we'll try to do a quick demo of this. So maybe there's some additional complexity that we could talk about and there might be some things that we could improve but if you right click here you can actually drill down in the level. So you have left click and right click if you do the right click. So now I would like to see, I'm looking at the national level. Now I would like to see all of the districts in this province for example. So then you can go drill down one level and it goes to that province and you can see the districts within it. And then you can go even further and go down to the facility level. You can then go back up to this level and then back up to the national level. So you do have that capability. It's maybe a little hidden because it's a right click but yeah, hopefully that's helpful. It might be something, so here we have this view profile capability maybe we would also be able to expose the drill down and drill up functionality potentially from there as well. Yeah, great. More than one disease on a single map. It should be able to have multiple layers. So you can do, let's say we want to do, this is A and C just as an example let's say we want to have HIV as well. I'm not sure this has, yes this does have data. So we can do a few different things here so first of all we can change the colors so that it looks a little bit different, right? So let's change it to something else entirely. Let's do this one. And then we can also change the opacity. So in this case we have two different layers, right? And so these are thematic layers which doesn't do that much when you have them on top of each other. You can turn one off and then turn the other one off so then you can turn them off on and off independently so you can see them on the same map. You could also create an indicator that combined these two if they were related in some way and then you could display that on the map. If you have events you can show both events together that could be done. But I don't know if that answers your question in terms of showing two different diseases or indicators on the same map. Yeah. Yeah, you can control the transparency. The transparency doesn't help too much in this case because you have two different values there. You could also, I mean the main way that I would recommend to do that is actually on the dashboard to have two separate maps, one with each and then you can put them next to each other because otherwise it's very difficult to combine the data in the same visualization without being a little bit confusing. There are some ways you could do it but it's a little bit challenging. Okay, maybe we move on to the next. Thank you very much for sharing. Maybe we move on to the next group. Regarding the visualization by map, in the Nepal we are using the 2.30 person so it is not updated but whatever in the system we are using the map as we can but in the lower level the health worker cannot be used because some difficulties there it is more technical than data visualizer and then pivot table. So we're using more in the central level by finding the different indicator at time. We are using it. So you're in version 2.30, correct? Yeah. In Nepal. So we don't have the proper complete geolocation of health facility but you are using the boundary only. Yeah. I mean there's been many changes since version 2.30 with the maps application. So I wonder if something like this would be accessible and usable enough for people at the facility level or if it's still too challenging? GIS and maps, it's a different kind of literacy that some people need to be trained on in order to understand what this means but it can be a very valuable tool especially at the facility or the local level for micro planning for being able to reach out to individual patients to see them on a map for example or households or those types of things. Do you think that this would be moving in that direction or is it still too complicated or too challenging for use at the facility level? Yeah, it is difficult to use. It doesn't have the practice of having a population of the settlement and the exact location of the health facility all over. So we use it in the, in lower level it's the municipality level only, not in the wild level and then community level due to some technicality. I see, yeah. Yeah, we use it for the micro level planning in the municipality level. Yeah, yeah. So that actually brings up a good question for the whole group, which I touched on briefly earlier today. One of the challenges we see quite often is just it's challenging to get accurate geographic information, right? So you need boundaries. You also need exact facility locations and facilities are changing all the time. Some of them are moving potentially. That can be a challenge, right? So I think that maybe that's a part of the challenge that you mentioned is the facility locations are not well known or not in, at least not available in DHS too. And so I wonder if there are other people who have experienced those challenges and maybe can share some ways that you've worked to address those. But thank you very much for sharing. If you have thoughts on that topic specifically feel free to share. Otherwise we can move to the next group. Sorry. No, it's great. Thank you. Okay, thank you very much. As during my presentation I mentioned that I'm working with Malaria Elimination Program so I can say something related to Malaria Elimination. Actually, we just started using this DSS to information system. And we started only during this month and it will be piloted for upcoming December up to December. And we have a dedicated MAS system for Malaria and we are using this for a long time. So from the next year we'll be using this DSS to instead of the existing MAS system that is a cloud based system. And this is the mapping using this DSS to is also still in the, planning for the next year, not now. So, but it does not mean that we're not using mapping. We are using mapping specifically at the central level. And very recently during this 2023 we started mapping using the GIS at the ground level. As you mentioned that this is not also easy because there is a difficulty in getting the boundaries and the health facilities, the movement of the population. As the country is now planning for the elimination so we need to have the detailed information at the ground. From where we are getting the cases whether it is from the at the community level or we are getting infection transmission at the forest level. So these are the things that has been incorporated in the GIS and we have a dedicated organizations also they are supporting us to do this kind of mapping. And we started from the highest malaria endemic areas. And hopefully by 2024 we'll be able to complete the mapping of all 13 districts, the malaria endemic districts. And maybe at the same time as you are planning also to introduce DSS-2 reporting system from 2024 that can be this mapping system that you already presented can be also utilized. But at present we are not using DSS-2 for the mapping purpose. Thank you. Maybe we have also the people from HIV. He can... Just to stay on malaria very quickly. I think malaria is a very, very key use case for GIS and for mapping as you've seen. Especially, I mean, you have spraying campaigns, for example, where you need to go and visit households. You have focus areas for specific mosquito bites. That I also missed because in the coming years we are also going to include this tracker for following up the case investigations and focus investigation and also response. That needs to be also incorporated in the system as well as if possible to include in the mapping. And this mapping will really help us at different levels, specifically at the central level. The managers at the central level can see the situation. The focus, which focus is going to be changed into the active to non-active, these kinds of things. And from where we are getting the malaria cases and what actions has been taken in that area. So this kind of information we are planning to include in the tracker as well as in the mapping. Yeah, so I just wanted to bring up this. This data is not very useful here exactly, but you have malaria cases and you have malaria foci for focus areas for malaria. And you actually can have a relationship between those as well. So you have tracked entities for people that are cases of malaria. And you also have tracked entities for focus areas and you can keep, yeah, basically track the relationship and actually map that as well. So that can be quite a useful functionality that was actually introduced in a collaboration with CHI specifically for malaria campaigns. And so there's a lot of functionality there. There's a lot more that we want to do as well. We're continuing to work with CHI and other partners as well to build out that functionality. We are also closely working with the Mahidolok Sport Research Unit. They are supporting us in doing this GIS mapping. Great. Thank you very much. You said HIV was another thing that you wanted to discuss. Not too much, actually. All are says Moshbi Guru Rahman Sir. In the HIV program, actually GIS is a great example in the data visualization, especially geographical location basis information. But in that time in our country, especially HIV program, not very much used to as a mapping because of field level stuff and program level stuff are not too much aware about the mapping. But day by day we will be used to the mapping and this is very important in our country, especially in HIV indicator. So HIV lots of indicator is international indicator and this indicator show is a mapping. It will be very much visualization to the all over the country. So we hope actually in the coming year we will be use the map is day by day increasing. This is from message HIV program. Thank you. Thank you. Any comments or questions for this group? No, okay. Thank you very much. We're gonna move on to the next group. Don't everyone jump up at once? Thank you so much. For Bhutan, I think we have been using this map for very simple purposes, like mapping of the indicator based things like in the cancer screening programs, we map the cancer screening coverage and also the prevalence of all this. And we also map only for the indicator based variables. Now, due to the indexing problems, I guess, we have not been able to use much of the map features. And also we have not been able to subscribe for this satellite imagery and OSM. That's why we have not been using it much, but we have been exporting this data, geolocation data based on the health facility and then using QGIS for mapping purposes. Yeah. Thank you. So you were a participant at the 2019 Mapping Academy, for example. Yes. And I think the fact that you're using in your country in Bhutan for the QGIS for a lot of the advanced analytics or even just analytics of maps is quite interesting. I wonder if you could talk a little bit for people who maybe don't know what QGIS is, kind of how does it serve your purpose? Is it useful, what can you do in QGIS? Maybe that you can't do in DHS too, or that you're using it for today? Yeah, QGIS actually is a very powerful mapping tool, topological mapping tool, and it's also free. So for the low resource setting, as well as developing countries, I think QGIS is the best, by the way, I'm not advertising for QGIS, but I think this is the best solution for us. In the DHS mapping tool, I think we have the features. I think many of the simple mappings can be done in DHS, and I'm very much sure that we will be able to do much more in DHS. But for QGIS, since we have the geolocation data, as well as when we need to add more data from the survey and also from the interview-based data collection data, and also some data from the other institutes and organizations like Climate Data and all, we have been using QGIS for all other purposes, as well as equity-providing and chloropleth maps, we have been using that. For malaria, I think we have done a projection once based on the data from DHS as well as other organization. So all in all, I think QGIS is a very powerful tool that everybody can use, and there are so many guides, videos on YouTube as well, which we can easily learn. And thanks to, I think, thanks to GIS Academy that I was able to use all this in my organization. Thank you. So I think that's very interesting. I mean, DHS2 mapping application has a lot of capabilities for displaying maps, for doing some analytics, for trying to investigate the data, and to make it actionable for people maybe at a lower level. But if you're trying to do very advanced analytics or to calculate driving times or to do all sorts of more advanced processing of the data, maybe with other data sources, climate data, trying to do mapping with elevation data in trying to figure out where steep slopes are for landslides or where flooding might occur, those types of things, that's something that isn't supported as a core feature of DHS2. But you can take the data out of DHS2 and put it into something like QGIS, which is free and open source, and lets you, you need a bit more understanding of GIS to be able to use that tool. So it's probably not something that anybody would use at a facility level or a district level potentially, but you can use that to really get some insights into what's going on in your country. And then you can even import that back as an external layer or as raw data into DHS2 and make it available to the district level, the facility level, where they're able to interpret it using the maps application or the dashboard if they don't even need to go to the maps application. So I think it's a very interesting use case. Thank you for sharing. Any questions for the group from Bhutan? Thank you for sharing. Thank you. So from Pakistan, we are using 2.40.1 version now for Airstream in malaria. Yes, we do using map. I have showed the maps in my presentation earlier. We are using maps and for data notification, but yes, we need to now go on to the data triangulation also and data analysis. And we felt a little bit of difficulty earlier that while showing the data notification on the maps, it has only three slots already available. Now I am seeing that there are a lot more than three slots. I mean data desegregation levels in low, high, medium. Earlier it was three, but now there are more than three so that we may be able to desegregate data more sort of an wider range. For data triangulation, we feel feeling a little bit problem to triangulate more than two indicators to be plotted at the same time in any map. And I think we need to learn more from the HIST Pakistan team on that. Yes, for GIS mapping, we are using with the support of KIT, APCON team from the Netherlands, Tropical Institute of Netherlands. And they are supporting us for GIS mapping for the hotspot identification for outreach activity for TV diagnosis and screening. So we are using those GIS mapping since 2019. Now I think we can now convert those GIS mapping onto the DHS too as well. Thank you. Thank you. And I was speaking with Adnan from HIST Pakistan as well earlier and we were talking about, I mean flooding is a big issue in Pakistan specifically and knowing the location of your facilities and the elevation where water flows, you can potentially have risk evaluation for facilities for flood risk as well as trying to do some detection even if you bring in more real-time climate data. I think that's a very interesting area of exploration. He walks in as I'm talking about him. But maybe you can talk a little bit more, I don't know if you wanna talk about the potential for maps in flood risk mitigation or risk mapping, Adnan? Yeah, I mean, I was just talking to Austin about a possible use case that we can use because if we see Pakistan, so it starts from very high mountains and then it goes down all towards the sea. So it has a very sloppy kind of altitude from top to bottom. And whenever the floods happen, they usually happen because there's a lot of rainfall and glaciers melting on the top mountains. And then all the water comes down towards the plains and then it flashes the whole country. So what we were discussing was that, I mean, with the extensions provided by Google, we can have these elevations built into the map and we can try and predict the way the flood is going to follow because it will not be going towards the high elevation areas, but it will be going towards the low elevation areas that we can basically predict and see that where the flood water will go, which health facilities and which population will be vulnerable. So this is something that we can do by, I mean, just configuring the DHS too and not using any external system. Yes, we can do many other, I would say, advanced tech, I think that can be used using DHS too. So one of the use case, yeah. And I think, I don't know if you were thinking in QGIS terms as well, how you would do that maybe because you have river valleys, for example, you know, if you know the volume of water, you can determine the risk level at different place. You import your facility locations from DHS too, then you can use QGIS basically to calculate a risk factor for each facility and then you could import that back into DHS too. So you could actually visualize the flood risk of different facilities within the system. So I think there's a lot of potential, there's many different use cases for GIS, not just within DHS too, but combined with some of the additional tools that are out there as well. Very interesting. Okay, yeah, move, yeah, go ahead. If any comments or questions, feel free. No? Go ahead. Would anyone else like to share? Yeah. I was just a comment maybe. And our friend from Bhutan also sort of mentioned in terms of use of QGIS, having more flexibility to depict multiple layers. I was just thinking that the example you presented there for ANC and HIV, let's say, having color shading on the same administrative boundaries. And a very common thing we'd be at times, tend to see in 2D printed out things at times as well for geographic mapping and depiction of indicators relates to one layer getting depicted through a color shading as the base layer. And then maybe the under indicator of ANC one coverage reflected as a bar across another added layer or maybe the density of a certain coverage through these sized bubbles and that sort of thing. So maybe that is something that could be considered within that as a feature you did. Mention and touch a bit on that. Yeah, I will do it just now. So this is actually something that is supported. That's a very good point. I didn't come up with it when I was doing this before, but if we have another thematic layer and we want to do HIV, for example. So my suggestion was that both can then simultaneously be viewed by the user on the same display. Yes. So here in this case, for example, you can visualize two types of data in the same map. In this case, maybe HIV and ANC, maybe they don't make sense, maybe they do, I don't know. But yeah, in this case, you have the bubble map for, it might not make sense if you're trying to visualize two different types of normalized data because as we said before, the normalized data or like per capita or per population, it's better to use the chloropleth maps. Bubble maps are a little bit better for raw numbers. But if you needed to, you could visualize both in the same way. And there might be ways that we could look into kind of layering of, it gets a bit too complicated, unfortunately, so one of the, you could do like shading with different types of bars or something like that on top, but it becomes quite difficult for someone to interpret what that is. So we try to keep things simple. But yes, that's a very good point, but it is possible to do some of these multiple types of visualizations on the same map. Do we have anyone else that would like to share? See some groups in the back being quiet. It's okay. Maldives, anything? Are you taking? Hi, hello, so we are from Maldives. So maybe this is not the problem with the maps app, but it's a problem with how Maldives is. So in Maldives also, we have configured the etol layer as well as the island layer in DHIS too. But when we try to look into one particular etol and try to see the distribution of a particular thing, maybe like how many children are registered in immunization, our islands are so tiny. And in each etol, we have so many uninhabited islands. It's very difficult to see them in one. It's very difficult to visualize them together. Maybe this may be a problem with other small island states where only 16% of our islands are inhabited. So that is something that we are dealing with, but usually if you just use the etol layer, it's a bit easier. But seeing all the islands in one map together, because there is a huge distance between one island and the other island is not so appealing or not so easy for us to use. So instead of seeing it like this, where you have each color right next to each other, you have one spot here and one spot here and it's very hard. And then maybe you have to zoom out so much that they're very small, so it's hard to see the color. So there could be an interesting way to approach that, which would be to use the alternate geometry feature. So you could actually create an alternate geometry for each of your physical etols or islands that is just a representation of where that is. So it actually fills up the space of the uninhabited islands around it so that it borders the other ones. So you could actually have something that looks like this. Each one of these might be a small island in reality, but it gives you a way to visualize that information. And then you could use that in the visualization to actually display it. So there's some ways you could work around that. And I think there are some interesting things that we could explore as well. I could connect you with our analytics and maps team to talk about some ways that we could support that. It's an interesting use case because I think it's something that a lot of people probably don't struggle with in their countries. Thank you. Thank you. Thank you. I'm representing FHI 360. We have a DHIS2 based aggregate system, which we are working across 40 countries, where we collect information for all the HIV interventions, including from outreach to care, all the cascade. So we use maps to map at the country level as well as at the district level. We have got data up to district level. So we actually map all of them at the district level at various indicators, including case finding, then initiation on treatment and all that stuff. That's one part. Second part, we have an issue with, again, somebody mentioned about point coordinates where our facilities are scattered. So we are trying to import the coordinates into the system and we are trying to map those health facilities as well so that we will be able to show the health facilities performance as well in the system. Along with that, we have around 18 countries where the individual trackers are being used, where we have data at facility level and we do hotspot mapping for identifying hotspots for key population. So while doing that, we use collect the coordinates and then those coordinates also are mapped on the geographies where these are congregated, which is used for our micro-planning. So this is extensively used in our area. So we want to go a little more deeper into the org units because we are specific org units at our lower levels, like DICs, data drop-in centers, and then other facilities which are operating specifically for key population. That we want to extend and get it extended. So that's what our experience has been. Thank you. Thank you. Does anyone have any comments or questions on that? I think facility point locations is a big topic area. I think we could spend quite a while on that. I'd be interested to dive in a little bit more. I wanted to show something as well. So I'm going to get that set up. But in the meantime, anyone else have comments or questions for this topic? Other experiences they would like to share? I'm going to try and get up a screenshot that may or may not work. So I did want to mention some other mapping functionality that I think is worth mentioning here as well. It's taking a little bit of a turn from analytics to data entry, actually. Because as we mentioned, it's quite one of the challenges. There are many challenges, of course. But one of the challenges is having accurate location information for facilities, for events as well, so that we actually have a feature in the Android capture application to use geolocation to collect points. So currently on the web, because you might not have a GPS on most laptops, for example, you don't have an easy way to kind of say, I am here. This is where this event is taking place. Record that information. But if you're on a mobile device and you're walking from house to house, for example, or you're driving between villages, you might want to collect the location when you record an event. And so we do support that as data entry. So you actually say current location, and it will give you the current location. And it will automatically fill that out based on your GPS. And just in the most recent version of DHS-2, we added support for measuring and giving feedback on the accuracy of the GPS location. So it will actually, previously, it would use the GPS location, but sometimes your GPS takes a long time to update and get you to the right location. And you're not really sure exactly where you are, so you just press OK. Now it'll actually give you an information say how confident is the phone that you are in this place. Let's wait until we have at least maybe a 10 meter or 100 meter accuracy level before we hit Submit. And so it'll give you that feedback, and then you can actually submit that. So it's a very good way to collect event locations when doing mobile data capture. The other feature of the mobile Android capture app product is that once you have that information or if you have any geographic points in the system, especially events, you can navigate to those places. So from your working list in the Android capture app where you have these are the households that I need to visit today, if those have locations associated with them, you can click and open Google Maps and go directly to that place. And so that's something that was highly requested feature for doing campaigns on the ground to actually be able to give, not just say, OK, I can look at a map and see where this is, but actually being able to say, this is the next patient or the next home that I need to visit today, I'm going to go directly to that. And I'm going to have a way to navigate there quickly. So that's a very useful functionality as well. You can enter geographic data on the web as well, but it's not using the GPS at this point. That is something I think we could add if that's available. A lot of laptops, as I mentioned, do not have a GPS, so it's not as useful as on the mobile device. But that is a way to enter event data or coordinate data. For org units, I think it'd be worth having a little bit of a discussion with the group here about ways that you've addressed challenges in boundary geometries or facility locations in your countries, because that is kind of fundamental to being able to use maps effectively is you need reliable and up-to-date information on where things actually are. So just to start off as a show of hands, who has had any challenges with having up-to-date facility locations or district boundaries in your countries? We have two. No one else. Wow, impressive. OK, three. Great. I don't know, do you have any insight into how you've addressed that or how you think it could be addressed? A lot of times it's political, but sometimes it's also technical in terms of how do you keep an up-to-date registry, make sure that it's synchronized between all the systems that are involved. Anyone have insight into that? I don't know if we have a microphone. Sure, do you have a mic? Yeah, I don't know if you want to start. Thank you. Well, we do have this agency in Pakistan that presents maps every year or every other year. Whenever you change, the district boundaries are changed. I mean, if you just take an example of Blotcheson, there have been many changes in the districts. Some districts have been merged. Some districts have been converted into two different districts. So this is one of the things that we can get mapping from there. Because doing it manually, it's challenging because we cannot see properly the boundaries and everything. So the best way is to just get the map from the government. As Adnan mentioned, about Pakistan's same situation, we are facing an IDSR as well. So what we are doing is we are in contact with Survey of Pakistan. They have up-to-date shapefiles. So that's it then. We are in contact with that. Yeah, updated shapefiles is definitely one thing. I think you sort of touched on it. But merging and splitting of org units of facilities of districts is complicated, not just for the maps. I mean, that's one thing. Obviously, you need to update the geometries and the locations, but also the data. What do you do with the data? That's actually something I demoed briefly on Wednesday as well, a feature that's coming soon is the ability to merge organization units. It's supported in the API right now with only two merge styles formats. But there's other complexities that you need to take into account. And obviously, it's helpful to have an interface for that as well. But that's definitely a challenge that we see quite a bit. Anyone else want to add? Yes? These are the real challenge in our country, specifically in hard to reach areas, the Hiltrak districts. And Mallory is also in the beginning of that area. So our plan is like that. We need to do the micro-statification at the BLS level. But the geolocation is available only up to the sub-district level. So that was the challenges faced during mapping of sub-districts and also at the BLS level. So what we did, we did it physically visiting the areas. And in that way, we're trying to finish the mapping of all the sub-districts in Hiltrak districts. But this is really a challenge. Yes, very much a challenge. And I don't know that there is one answer. Probably many things that we can do, but it's a challenge for sure. OK, I wanted to share a little bit. Does anyone else want to comment on that topic or discuss ideas? Crazy ideas are very welcome here. So safe space. I did want to share also some. We talked a lot about the Google Earth Engine support here. So I wanted to demonstrate that quickly. I'm going to change this back so that I can see it on my screen because it's very difficult to do down there. OK, so I'm going to talk a little bit about the Google Earth Engine capabilities that we have in DHS2. As you can see, we have many different layers here. Population building footprints is a relatively new one, but that's quite useful. So let's say let's do a building footprint layer. Let's go ahead and add this layer here. It takes a minute because it is actually going to Google to calculate the building footprints and then get those back, and hopefully it will finish here quickly. So this is loading basically the satellite determined all of the buildings, all of the structures in the country of Sierra Leone. So if we start, I'm not going to need to dive in there too much, but if we start to zoom in here and we get closer, we'll see that we get more and more detail. This is a city, obviously, so there's a lot of buildings there. But let's try a more rural area, maybe this here. And I think this is quite an interesting layer and very useful in spraying campaigns, for example, but many other use cases as well. To be able to determine, I mean, this is obviously a town with a number of buildings. You can see exactly where those buildings are. You can actually use this to plan a campaign of spraying. You can use this to determine where populations might be that you might be underserved by your health system. So that's a very useful functionality. We can also add, let's say, population by age group. So as we mentioned, we can split this up. So let's say we want males and females under five years old. Let's see all three of these, not both of those. So let's just say years one to four. So we have both males and females under four or under five. And we're going to load this layer as well. So this is an interesting use case, because you can actually see the two different layers, two different colors here. So we have the building footprints, but we also have estimates in those 100 meter by 100 meter grids of, let's go to this same place again, of the population and how it is, how many people are estimated to be living in that little village who are under five years old. And so we can actually see that in this specific point, sorry, this is the entire region, because we aggregate by the region. We can choose how we want to do that. But we can actually see show the population age groups just in this hectare or 100 meter by 100 meter square. And we can see that the estimate here is that there are eight people in the 100 meter by 100 meter square who are under the age of five. So there's maybe one or two families that are living there. So this is really useful also, and we can go back to the example maps that we had here to show another example from South Africa. And this one was actually where they used the population layer and the building footprints layer as well to identify rural populations that had no access to health facilities that they didn't even know existed. So they were not registered at all in the system, because it's a very, very remote location. So they were able to use this building footprint layer, this population layer from Google Earth Engine fed by WorldPop, to find people with no services at all and then can do an outreach campaign to that community specifically. So this is a really interesting use of just being able to put this data on a map and explore it. As we did just there, we dove into a specific village, and we could learn a lot about that village just by turning on two layers that were very easy to configure. There were some other examples of that as well in this workshop. There's a lot of useful examples here. Let's go ahead and actually let's do that one. I don't know if this is an interesting exercise, so please stop me if it's not. But can someone volunteer maybe to interpret this map for us? Say what story is it telling us? And why is it important? Why is it useful? Any volunteers? Maybe people don't like this exercise. That's OK. No volunteers? Yeah, sure. Thank you. So it's basically showing red in population. Red color basically is showing population and the confirmed cases of TB. It's actually showing where the population is higher. The cases of TB are also. But in some areas, maybe the population is low. That's why the cases are maybe zero or low. So this is. Yeah, it seems that way. It looks like there's some interesting case down here where we have relatively higher density of population, but not as many cases of TB. So maybe those are not served by a health facility. Maybe that's undiagnosed TB potentially. There could be a lot of things to investigate with just looking at this map. So yeah, it's very interesting to look at visualizations like this and start to see what you might. Yeah, what you could learn from something like this. But also it does not say why in some areas the number of cases are high. It's difficult to say from seeing this map. Yes. In that case, we need to go to the data. But it gives you a place to look. This is where you could start looking at. There's a high population and a high case of TB here. We could investigate that in the data. Maybe the data doesn't tell you what it is. So you need to go and actually visit the facility and do some investigation. So yeah, it doesn't tell the whole story. And I think that's almost all data will not tell you the whole story. And so you need to use it as kind of a way to spark curiosity and figure out what's actually going on. That's a very good point. Let's do one more. Well, the country is like South Africa, which is a high burden for TB as well as high burden for HIV. I think we should triangulate the HIV TB cases. I mean, TB case notification against the HIV notification in the same district. So that we may be able to know that there could be a co-morbid cases there instead of TB alone. So perhaps we should take a lot of any epidemiological assumptions while plotting the maps. Not only this only shows the notification details. It doesn't show anything else, which might, I mean, any confounder, which is basically affecting the notification here. So it is a simple correlation that the population is high, so the notification is high. But we need to know something more about that. Yeah. It'd be interesting to know also if there were higher numbers of suspected cases in some places than in others. And co-morbidity is a very, very important one. Again, you need to find some way to visualize that on a map or to explore it in another way. Very good point. Last one. And again, I haven't seen these before other than at a glance. So I'm interpreting them with you. This one, the legend is a little bit. So this is actually a good example of maybe, what does 100 to 250 coverage mean? I'm not sure in this case. So we have 200% coverage in some cases, which feels like maybe that's an indication of something that is not going right or in the data collection. Probably your denominator is not correct. So it's an interesting case where, again, accurate population data. We had somebody mention that if you have a census from 2004, population has grown a lot since 2004. And if you're using that as the basis for your denominator for immunization coverage, you're going to have great immunization coverage, but it's not actually going to be covering people. It's going to be covering the 2004 population. And so I think that's another big challenge. In addition to the location of facilities, making sure that's up to date, is having accurate denominators. And this is something that's a big problem. I wish maybe Jorn or Ula were here to talk a little bit about this as well. But the denominator problem is a big one. And it's also a challenge sometimes politically, because the person at the facility level, they just want to know how many people are in their district. They want to know that so that they can provide coverage. But sometimes the political motivation is to present a number that looks good. And that can be a conflicting interest. So what's interesting about DHS too is that you can have both. You can have the official census numbers, for example. And you can have more nuanced, detailed estimates at the facility level so that people can try to find the people that are unreached, for example. But it's a big challenge in a lot of places. Any other, anyone else want to interpret or comment on this map here, what it might say? Yeah? Thanks so much. I think the map actually shows the MRA coverage. MR1 coverage, sorry. And similar to this, we also experienced this, the coverage going more than 100%. So it typically, for me, I think it typically depends on why you are plotting that map, the typical use cases, the purpose of the map. For example, in my context, I wanted to see the coverage as well as the use of the vaccine doses. So in that, for the overall national or at the district level, it will be 100%. But at the sub-district level, it will go more than 100% because district A, the catchment population of the district A can receive a vaccine from district B. So it makes sense in that way. So I think it also depends on how you interpret the data and where and the purpose that you plot that map. Absolutely, yeah, definitely. Thank you. Any other comments, thoughts on this particular map? We're going to move on to one more. So this is two different maps, obviously. But it's telling you an interesting story as well, I think. Anybody want to venture a guess as to why it's interesting to have both of these next to each other? So if you show someone this map just on the left, what are they going to assume? The entire south of the country is bad, right? But that's a big area that needs to be completely overhauled with their help system in order to address this. But if you go a little bit deeper and you zoom into the next level, you'll see that it's not the whole south. There's very good coverage, actually, in some of those districts. But you need to go a little bit deeper and find the next level of that data. So it's important that this is a good example of the map tells a story, but it can tell the wrong story very easily. So if you show someone this map, they will have a very different assumption than if you show them this map, which gives a more detailed action plan for how to actually address this problem by maybe going to the east or the west instead of to the south of this country. Any other comments or thoughts on this one? Has anyone ever seen a map that was misleading before? I have. Many times. I think that's a very interesting point as well, is it can tell a story, and sometimes it can tell the wrong story. And it's important to think about what story you want to tell and how to make it actually actionable and useful. This one is a little bit busy, but can someone venture a guess as to what story this might be telling? It's showing a pentagon number of values on the legend, if we can see. And it also shows the river, roads, and the health facilities, and also it is showing the population as well. So the color intensity is showing the pentagon number value, where it is high. And clearly, as I mentioned before, we can see the river side. We can see the roads, and health facilities, and number of density of population on the map as well. And I don't know the last one, which is showing the area in kilometers. Sure. So does anyone have a, and that's correct. That's all correct. But what story is it telling? What do we notice? I mean, there are many things we could notice. I think there's a lot of population, but they don't have roads to access the health facilities. That's why in this part, they don't have so much good vaccination, I can say. I think that's a good observation. So we have a lot of black dots here. Black dots is population. A lot of black dots that are not covered. So I think these circles, it's not labeled here. But I think these are one kilometer circles around the facility. So the white circles, I think are, yeah. So you have basically. It's a catchment population of facilities. Yeah, so there's a lot of people that are living more than five kilometers. I think it's five or 10, maybe 10 kilometers from a health facility. And there's no road. Only going to have health facilities in their areas. Exactly. And then you have poor vaccination coverage. And that could be a cause for that. And it could be something you could address to try to address the poor coverage. OK, I think that's it. We have about 15 minutes left. Just want to turn it over to all of you to ask any questions that you have or anything further that you'd like to discuss about maps, about GIS, about DHS-2 in general, about how to implement programs in your countries using mapping data. This is the same quiz as before. So you don't need to. You can fill it out if you haven't filled it out already. Very helpful. But does anyone have any questions or something that came to mind that they would like to discuss? Can you please explain the land cover layer? Sure. So to explain land cover, I'm going to add this layer. And I'll actually get rid of these other two. And let's leave building footprints, actually. I think that's a good one. Put that one on top. So as you can see here on the left side, we have different types of land coverage. So this is using satellite data to estimate what that land is used for. So is it cropland? Is it forest? Is it urban development? Is it what type of forest is it? Is it a grassland? Is it snow? There's lots of options there. Is it water, of course? So we can start here on the left side. So we can start here on the border. And we can see what this looks like. So we have some islands that are not really islands. It looks like they're wetland. So this is the wetland color, which means that it's probably like a marsh or some kind. We can go a little bit further in. We see we have some more grassland, savanna, a little bit of wooded area. We might have some cropland somewhere around here. Let's see if we can find some. Yeah, so this is a cropland slash natural vegetation mosaic. We can see maybe where we have a little bit more grassland. We can definitely see where we have more urban environment. So if we look at urban, we have that. Where we see urban. Urban is red. So we can see here we have a city, for example. And so this is urban development. It looks like it's surrounded by cropland or maybe a forest. I need to look into that a little bit more. And you can actually look at this closely and say, just in this point, this is cropland natural vegetation mosaic. So this can be useful for a number of different reasons. I mean, it can be helpful for targeting outreach. It can be helpful if you're doing investigation into standing water for malaria, into water areas for potential flood risk. Probably wetlands are particularly relevant there, but also maybe different types of investigation. It can become very valuable in a number of other cases when you're trying to figure out the intersection of health and climate, health and environment, one health, those types of things. Does that explain, answer your question of what it does? There's a lot more that you could do with that that I'm no expert on, but there's a lot of interesting use cases, I think, for having this information available. Any other questions on this later time? OK. Any other questions about, yeah? Just two quick questions. Number one, I think for my organization, we were not able to use, really use, DHS to map is because of the legend, the interval of the legend. So it auto-generates based on the number of the intervals, the categories that we require, for example, and is there a way to customize it and then specify the intervals or the bin? The number two is a few notes on how to subscribe for the satellite and the thing. Yeah, good question. So the first one, are you talking about a thematic or chloropleth layer? Yes. So in this example, we have a chloropleth layer for ANC coverage that I'm going to add. In the Style tab, you can use either an automatic color legend or a predefined color legend. For automatic, you can specify either equal intervals or equal counts. So this means if you have one, two, three, four, or five, you can either say, well, one to 10. Then you can either say one and two, or you can say the first 10, the second 10, the third 10, different ways to split that up. You can have up to nine classes that are automatically generated. So let's say we wanted nine different levels of gradation. And then this will, because we're doing equal interval here, you'll see it automatically picks the minimum value, the maximum value, and then splits that up by about 10, 10.5, 10.4. So each of these is about 10.4 wide in terms of value. If we chose instead equal counts, it would do it based on the actual values that are in there and say that we're splitting this up evenly based on kind of how the data is distributed. So we have roughly one, it can't do exactly, but one or two count values in each of these locations. You can also do, as I show here, you can also specify a predefined legend. So this one is ANC coverage, which is defined specifically for ANC, you can do that. That is defined in the maintenance application. And there you have a bit more flexibility in defining exactly what that legend can show. Let's take a moment. So we have legend here. So this ANC coverage legend. And there you can define exactly what the start value is, what the end value is, what the color is, and what the name, the label is for each of the values in your legend. And so you can actually define a custom legend as you'd like. That answer your question about legends? Yes, thank you, thank you. And then your second question was about how do you sign up for Google Earth Engine or get access to that? That is just here. Let's go back to this one here. Should be, yep, there we go. So we'll share this as well, but you can search on docs.dhs2.org. There's documentation there to show you exactly how to access that Google Earth Engine layer. Basically just send an email to maps.dhs2.org and we'll get you signed up. So we used to be, every country had to talk with Google and then they would have to validate with us that it was using dhs2 and was an actual country. But basically you just can send us an email, we can generate you a key and you can go from there. So that's as basically as simple as it can get. There's some additional instructions that you need to actually install the key in your system. So you have to put it on your file system and start dhs2, point to it in your dhsconf file. But it's as simple as that to get signed up. Thank you for the question. Any other questions? We have a few minutes left and then we'll wrap up. I know people are getting towards the end of the last day. Oh no, that's okay. Any other questions, thoughts, comments, jokes? Do they have a joke? All right, with that I think we'll wrap up maybe a few minutes early. I don't know if you had anything you wanted to talk about or add or no. Okay, thank you all very much for your time. Hopefully it was a good session. Thank you. And I did very little talking, I'm happy to say. So I'm glad that we had a discussion and thank you all for sharing because that's really what this is about. So thank you very much. Right, so we are waiting for the session downstairs to finish and they all join us for the closing ceremony. But until we get started, those of again, related to the last announcement I did. So those of you who are sponsored from the Hispacia Hub, you can collect your visa fee from the registration desk. So please, those of you who are sponsored from the Hispacia Hub, your visa fee can be collected from the front desk. And also please note that the visa fee will be paid in Sri Lankan rupees because in Sri Lanka we are not supposed to hand over any cash other than Sri Lankan rupees. So unfortunately we are not able to give you money in US dollars. So you can get it changed at the airport. As an entity, we are not allowed to disperse any USD payments in Sri Lanka. That's why. Thank you.