 and welcome back to TV UP Health Issues. This is your host, Dr. Teddy Herbosa with our guests. This is part two of the UP COVID-19 pandemic response team. And on this episode, we will be discussing about data, data management, data analytics and information. With us today is Professor Mahar Lagmai, Dean Jomar Rabahante and Professor Peter Kaitot. Maybe we'll start with asking Professor Mahar. Can you tell us about the product of the UP COVID-19 pandemic response teams dashboard that it created on the web portal? That's right. We showcase all of the outputs of the efforts of the UP pandemic response team of the entire UP system. Those who wanted to help in the COVID-19 efforts in a website called mcov.ph. Now, the people are seeing it now on the screen. Yeah, that's spelled E-N-D-C-O-V.ph. And it contains a lot of information. It started off as a website that was supposed to cater to UP employees, UP faculty, employees, alumni, students. Because we're a big population as well. But eventually, when we put it out, a lot of non-UP people also were interested because it contained information that I believe were helpful to them. At least that is the feedback that I'm getting. At one point in time, I think it had several thousand users per minute. Depends on the season, depends on the time. If the COVID numbers or active cases are low in the day, then yung visitors also go down. But quite recently, when there was a search, we had a lot of downtime of the mcov.ph because there were a lot of visitors using the website mcov.ph. Now, the mcov.ph, again, is the showcase for the outputs of all of the research that the UP faculty and volunteers of the UP pandemic response team are doing. In short, I'd like to state that this kind of output showcases not only the current data. It showcases yung mga now cast would probably one day or two days. And then also the forecasts, risk assessments as well as scenarios of the COVID-19 problem that we have. There are inputs coming from mathematicians, statisticians, geographers. We always talk through the Facebook group and generate the output that you would see here. It even includes a part there which is not really artificial intelligence. Some might call it as AI already, but in the true sense, I believe that it's not really AI. But some may consider it as AI already and it's called YANI. There are also non-mathematical or non-mathematical aspects here in this website. Like the work of the group from NC Pag, Chris Berset where he compiled all of the policies, policy statements, advisories, policies of the different LGUs, et cetera. So basically what I'm saying is that it contains a lot of information, a lot of infographics, policies, information that people would need including the now casts, the forecasts so that we can anticipate and scenarios so that we can plan in advance our fight against COVID. So to show you the details of what is inside the ENCOV.PH, I would transfer the microphone or the floor to either Peter or Jomar. Go ahead, guys. I'll just butt in when necessary but I'll leave it to you, Peter and Jomar. So for me, for the non-mathematician in this particular panel, these were data visualizations. To me, I got to see how the numbers look like because if you just showed me numbers, it doesn't excite me as much as it excites Peter. Because they are presented in visual format, I'm a surgeon. I'm trained to look at stuff and visualize stuff. So if I see a lump, it's there, it's a lump. So this is what it has done for me as an individual that's a non-scientist looking at this table. So maybe who will describe it? Peter or Jomar? Peter. Let me describe first the features of the dashboard and the maps and I'll give it to Jomar to talk about the projections and the city versus COVID. So first of all, this is the case overview and the first thing that you'll see when you access the ENCOV.PH dashboard. Well, here we have details for NCR, the total cases graph in terms of cumulative counts. The current existing counts, though these will be updated later on, these are still reflecting yesterday's counts, but they will be updated later on today. And also some breakdowns of the cases. Now, there is a second tab here, which is PSPHP graphs. And this is work that Jomar, Darwin, and I do with the Leads for Health Security and Resilience Consortium, which is formed by the PSPHP and its members. And here we are showcasing a lot of the products that we've done. For example, the national epidemic indicators for time varying are case fatality, recovery rate, and you can also in fact interact with this to look at the daily cases and active cases and the growth of cases through time. And you can also change how long you want to look at the data. And as we've talked about in the previous episode, the look on the RT. I believe there are still updates happen because there was data release earlier at 4 p.m. So these will start to update and this is now okay. So we have the current trend of the RT that we're seeing here. It's slightly going down. And I anticipate that as the new data comes in for April 7, it will again go down because of the current reported cases by the DOH. Now let's go with the map view. And this has gained a lot of exposure recently. Our map statistics at the default, you'll see the municipal counts of COVID-19 color coded so that you can see in terms of the different counts per municipality in the Philippines. You can adjust the data that's presented on the map by looking at the data layers available. And we do have, for example, public schools in DepEd because it has been, the recent talks about, there have been talks about opening schools on a limited fashion. However, because of the growing number of cases, there have been a pause in terms of that policy. There's also recently a more helpful tool in terms of showing the hospital, hospitals in NCR and most regions in terms of the capacity that's still available. And if we, for example, zoom in further, we could access the existing COVID-19 dedicated capacity rate 100% in Manila Medical Center as of April 4. So here, at least with the available data that is given by the DOH, we are able to at least look into what could be the available spaces for the different hospitals in NCR and also outside NCR as well. Right now, there's also data on COVID-19 testing centers, though this still is for updating, especially that based on the most recent count that's available in this map, there is about 100 to 150 laboratories already registered in address here. Later on, the additional laboratories that have been reporting recently, around at least 200, would be now integrated in this map as well. Now, there's also the quarantine facilities map. The COVID hospitals, they're already redundant with regards to the details with the occupancy rate, so those are already okay. Now, let's look at the active cases in terms of we're monitoring active cases here. There's the provincial level active cases. So as you can see, with much of the nationwide, we can see that there is the greater Manila area having a lot of cases, but as you can see, it also kind of bleeds over to Pampanga, Patangas, and there have been certain noted high cases, high number of active cases in Isabela, still in Cebu, that could be helpful in terms of showing those provincial active cases that are existing. This one, we are due to update the active cases per 100,000 population in the municipal level, and in the future, we are planning to do this at the barangay level as well in terms of giving that information on the concentration of active cases per 100,000. More recently, we are able to update it for Metro Manila because this is the kind of data set that's available right now by the researchers of the UP COVID-19 pandemic response team. Blue is not a good color because it means that almost all of the barangay have at least 100 or more active cases per 100,000 population. You might see some areas of which that would not be the case, like, for example, much of Tagig, a lot of northern Kaloakan, some areas in southern Kaloakan, and the southeast portion of Makati. But as you can see, it is a large distribution of a large density of cases in much of the barangays in Metro Manila. More recently, we've gained some popularity with the public in terms of showing the dot maps. This has been a very visceral data display. A very dramatic data display in terms of really showing the distribution and the scale of the epidemic. More recently, we've been releasing maps, not in ENCOV, but in the future, it will be available in ENCOV. We've been releasing maps of different areas outside NCR, like, for example, Region 3 and 4A. And we've also shown an initial map for the whole country. But I am part and we are looking at the mapping activities of the UP COVID-19 response team. And we are going to be developing more interesting and interactive maps for the public to interact with regards to the active case spread of COVID-19. In one of the policy papers that the UP COVID-19 has written, we were also showing about the risk assessment per province based on the existing number of active cases. This is where what we call the probability of outbreak. And this is based on research looking at how the concentration or density of people in a province and later on in a municipality as I will show later. And the existing number of active cases can relate to what is the outbreak risk of a province in terms of spreading COVID-19 to the point that there is a risk that it might not be easily controllable now. So here we can see that the 99% and above that means that there's a probability of outbreak that is at least 99% means that there is a very high risk that COVID-19 will spread. And in some regions it has already been an outbreak of COVID-19. There are some exemptions that we will see here like for example, Sulu. We also have here Davao Occidental. And if we look at another level of the probability of outbreak which is the risk assessment of outbreak and this is for the municipal assessment. And with the municipal assessment a lot of these are mostly like very sensitive because if we see that there's at least one case in a municipality especially with the kind of features COVID-19 has right now it could be a risk for that municipality in terms of an outbreak of COVID-19. So this is the kind of data, the PC in terms of the municipal level. And in fact with the kind of data that we have here you could see more than one data layer but that is for the interactivity that we have here of the map with the COVID-19. Before you proceed Peter, let's just be clear this format, this dashboard is free to the public. Yes, this is available to anyone who can type who has internet access as a web browser and can type in ncov.ph and they will be viewing all of this information that you were just showing. Correct, yes pa. They don't have to pay, they don't have to log in, they don't have to have a subscription etc. It is freely available you just need to have access to the internet and you will be able to see these information available to the public as a service of the university to the people. And what Mahar promotes as data. Can you show some more of the other features of this Jomar? You also have some of your model projections and models in graphs and charts that have been very useful to the policy makers government officials and ordinary people. Can you show what those features are? In the projections tab in this projections tab we have three types of projections. The first one is the in terms of cumulative number of cases and then the second one is the gets sparred cases. And then the last one is on the epidemic curse. The new cases per day. So I think we can start first with the new cases per day. This one as you can see we have arranged here of possible levels of cases in a day because we run 1,000 simulation scenarios. The green one is the optimistic side where all of our interventions will be okay, will be effective and then the red one is the pessimistic scenario where most of our interventions are not working. And currently we are at the middle of the red and the orange. The orange is the mean. However sometimes there will be some peaks like remember we had 15,000 cases in a day but here of course it will go up to 15 but there's no 15k here yet. I need to update the actual numbers there but it's very hard to really plat there the actual because we know the actuals are dependent on the validation delays of the OH. Validation and reporting delays of the laboratories and all the technicians and people getting sick and not reporting for one week. But in essence I believe if we're going to delete those delays it's really somewhere at the middle of the red and the orange. So that we still need to do more to make the number of cases lower and lower or in the language of Peter it should be way lower than RT equal to 1. It should be less than 1. Then in terms of our projections about the debts debt here it is somehow along that side the blue one na jen yung However I need to point this out if you go at the 2020 2020 period you can see we did the projections of this debt around December 25 last year and we used the data for fitting the blue one. It's the actual reported debts during that time December 25 every day there will be additional debts to be reported so the actual debt now is the black one. So it means we really delayed in the reporting of debts but still with projection it's along the projected range and then for the cumulative for the cumulative here we actually revised our simulations here because previously our cumulative is along that band yung ano ba yan it's like a greenish brown and up to the red but we revise it because we now we experience now a surge in cases so it is something like that not so good increasing it actually looks like a line but actually it's exponential so I hope that after our interventions we're not really going to that level so this is in terms of projection that might help our decision makers in their response and then the second the other tab is the city versus COVID this is created by the group of of course the UP COVID-19 pandemic response team headed by Sir Chris Berset is led by Sir Chris and so here is actually I really like this tab because we know most of the cases are in cities and we really need to monitor the cities so dito what we can see are the active cases recovered and expired cases in cities and of course compared to the whole country and from here we see it's 71% coming from the cities for the another favorite is the road to recovery so this is also one of my favorites okay so here we can see the different cities we can look at the data okay it just takes a while because we have we have a lot of data it can just take a mean and data is that you guys have analyzed the data and it comes out as process information unlike our department of health that keeps reporting everyday is static numbers so what I love with you scientists and mathematicians is you visualize it, you've gotten that data our data sources is still the department of health right we get the raw data but you put it and you crunch it and you analyze it into maps into figures and graphs that become useful so that's the part that's not happening on the department of health side is that correct? can I answer that? can I answer that? go ahead Mar the model of the pandemic response team is to get the information out because the people need it they need to visualize what is happening and it needs to be on a nationwide scale and the motivation really is that we need to deliver through digital technologies we have to maximize digital technology to our advantage if we don't do that we are not maximizing where the losing end the information that is provided here is not only static or current data it's got a lot of analysis already crunch numbers the numbers have been crunched by the mathematicians and they're all out there seen in ncov.ph nationwide so that the LGUs can take a look at them and be able to use them as their basis for decision making I think that kind of model needs to be adopted because the model now for a lot of information like this there's a big discussion about it some say that it will generate panic it will create a lot of trouble some people may use those data no what is important is that people know the real score people need to know the truth what is happening and people need to know the analysis because they need it because after all everybody is affected our loved ones are affected and prof mahar you always talk probabilistic approach to DRR so you always talk about static number but the probabilities that happen out of the risk that is occurring or the hazard that is occurring can you explain probabilistic versus static type of analytics well I can explain that in simple terms we have a statistician here who can explain it much better I defer to Peter to explain the difference between deterministic and probabilistic probabilistic so let's educate the people well for one with regards to disaster risk in statistics when we think about risk built into it is the uncertainty and it is by planning with regards to uncertainty that is important with disaster risk reduction and management I personally in terms of my research in economics and finance I've always been pondering what is the worst thing that could happen to our economy what's the worst thing that could happen with financial markets that's why much of my personal research before with the pandemic response team is about talking about risk management in terms of finances and always planning at what could be possible scenarios so that even when these occurrences happen we are at least padded in terms of the impact of these uncertain low probability but high impact scenarios wonderful so you guys are not only disaster scientists you're also what I call data scientists because you're crunching the data analyzing the data information for policy makers that's correct right so tell me the story and any of the three can actually jump in we had a lot of struggles with how we manage our sources of data how we were you know validity of the data the quality of the data we even put out one policy paper on quality of data I remember so let's talk about that because these things improving data management would actually improve our response to the next pandemic anybody want to share experiences siguro simulan ko na po muna about this especially with regards to the context of our policy paper number 6 which is about with the data management the issues for me that really started off as a rant to how I was looking at the DOH data in terms of them changing the format I have to tell you guys remember I have to tell down that policy paper anyway go ahead tell the story and it's the ideas like why is there these discrepancies between different sources there have been like sudden changes in terms of the missing of the missing data and then suddenly it changes the status there was a change in the magnitude as well somewhere in the middle right yes they did they were inconsistent with date formats and everything it almost changed every one or two weeks which really is difficult if you're already established into analyzing the data so that when it comes in again you could process it quite quickly so every time we've always been like so you guys always told me the data was dirty very dirty you said the data was very dirty and you had difficulty analyzing and crunching the numbers because the data you got the data dropped every day was dirty data that's right yes it is like you would think that Manila is a city in NCR but then suddenly there would be another data of Manila somewhere else correct correct and suddenly we would be finding like barangayos of Manila being declared as cities or provinces like for example even your place wasn't found yes which I slightly felt why is my town missing because I always religious in terms of looking at the situation of my hometown I've been pretty much sharing that and blurting that out to really point out that there's a problem on the data there was another problem we had actually at the beginning when we established ngov.ph as a website there was another group that actually created a facebook site named ngov also and actually opposed our being published as an ngov but I know about that because later on after opposing later on they approached us and asked us to include their products into the ngov site on the basis that the ngov.ph was I think they thought was more popular than the site so we had more visitors than the site that they have we welcome them all efforts are welcome inclusive we have inclusive approach to this pandemic anyone who wants to volunteer was accepted that's correct but ebps ang ang probabilistic it's really about making scenarios and I would like to put an international flavor to this one IPCC Intergovernmental Panel on Climate Change when we talk about scenarios tamay sinabi we're trying to deal with uncertainties and they would say at IPCC that the goal of working with scenarios is not to predict the future but to better understand uncertainties and alternative futures in order to consider how different decisions or options may be under a wide range of possible scenarios so that is what we are dealing with because we have found out in disasters that have been happening here in the Philippines we always hear the people saying ay, first time na nyari ito this has never happened before we've never seen a flood until we solve that problem we will always be hearing it we have to prepare for scenarios that are bigger than the historical record or bigger than what we have experienced by doing that we are doing anticipation we are planning in an anticipatory manner and by doing that we are preparing better for future events and that's the kind of risk that we have to prepare for depending on our resources now ang importante probabilistic marami po tayong nilalagay dyan so that the LGUs can use it for their city or municipality but still talking about the international thing ang total na rinaman tayo that is for disaster risk reduction ang concept ng open data is a trend already in disaster risk reduction europe, america they have been sharing all of their data sets because they know that by sharing more people can work on the problem more scientists can generate knowledge that we can use now there's a lot of data that is just before you go ahead let me just explain what really happened because we asked for a data sharing agreement with the department of health remember that? we had a data sharing agreement and it took like 3 months or 4 months and it wasn't moving data sharing agreement so that's another problem we encountered because they were claiming data privacy act and they didn't want to share because we wanted to go all the way to the barangay and they were claiming their data privacy officer was claiming we cannot just freely share this but we were on the other end saying this is open data and you can anonymize the data and share it to us but i think nothing ever happened to that data sharing agreement at least we tried but there are stumbling blocks what we really need to do is to push this because it is the trend for addressing our disaster problems correctly we need more scientists science needs to be trusted Peter was talking about yung data na hindi maganda that can be improved he was also talking about people other than people in government who work on the data so kapag ka naglalabas sa science yung nilalabas needs to be replicable that is a foundation of science your methods whatever you say needs to be replicated can be replicated how can you replicate it if there is no data that is available for you to review and check the methods so important if we are to trust in our decisions and if people are saying that it is science based the science must be amenable for review and to review it you need data so data must be open and on another aspect another aspect can you hear me? on another aspect when you communicate like now i'm talking to you and we wanted to when we communicate somehow our intent is to be able to convince but how can you convince if there is no trust it is very difficult to communicate without trust also to work as a whole of society we need to get all of these data sets in silos so that we can do a better analysis kasi when we plan against disasters we plan it not just for the health sector we plan it for agriculture we plan it for education tourism, energy infrastructure and all sorts of other sectors for the whole of society correct so the more data that we have the more data that we can integrate through open data not residing in silos the better our country will perform against disasters if we are able to perform better against disasters progress can take place development can take place unhampered by all of these problems about natural hazards wonderful, wonderful data we actually started talking about a dashboard just a plain and simple dashboard a lot of data and lots and lots of analysis and graphs and visualization and from there we've learned so much about this pandemic which was new the way it behaved, the way it spread even how our data was managed even how data was collected by our own government colleagues in government agencies and I think you guys have been able to show the value of what the academician can bring to the table in the time of a pandemic I'm sorry guys, this has been a very interesting discussion but I guess we're also short of time but I'll ask for your final words and I'll start with Peter Jomar and then Mahar for your lessons learned here in this experience on data management in the time of an epidemic or in the time of a pandemic Peter? So with regards to all of the data that we've contained in ENCOVE and also with our maps there's one thing for me that's always crucial to reflect upon the numbers the statistics, the dots they're not just these esoteric items these are people and it is by having the data being made available to all that we can start giving more context to the situation by looking that these numbers, these dots these figures, these are all people and we want to tell the story of people for the improvement of people's lives Wonderful, the power of data to tell stories I know of epidemiologist who does that very well he tells stories through these graphs and his pictures and I think that's what you guys are doing you are storytellers who are telling that data John Mark? I think in general there's no perfect data set especially big ones huge data sets are not really that clean but I hope it will be our target to make it near perfect and I hope DOHB4 I hope they appreciated what we did last year in pointing out their errors we don't want our data sets to be as what Peter is saying to be as clean as possible to be somehow near perfect because we are talking about lives of people and I think the other take home message here is we also need to look at different perspectives about data sets if we think there are some issues with the national data we can also work with LGUs because they have their own data set and we might look at their own data set so we can have different perspectives different like faces of the story so yun po, thank you thank you John Mark and finally Mahar we are talking about the power of data especially during times of emergencies or times of disaster you don't really know what kind of information or data sets that you may need we don't know what we need for example during a landslide everything was covered the size of the landslide was 3 fourths of UP and they were buried under 30 meters of rubble and there was a text message of a teacher said that there are 200 children here there were 2000 people searching rescue people trying to unearth them but they did not know the location at that time they needed a GPS point everybody knows what a GPS point is but that data set that GPS point was not available as open data it took 7 days before that data set which was held in an institution to be given to the searching rescue people by that time it was already rose and there was no chance for the survival of people so during disasters open data is really critical it's very important but it's not only during disasters that we need them we need them to plan our local government units well not just in one sector but across all sectors so all data sets in silos must be integrated so that we can have good plans for all 1700 plus cities and municipalities of the country it encourages or fosters trust it fosters transparency that is required by science and that is required in communicating disaster information so open data and the rest of the scientists in the world believe is the way to go in effective disaster reduction and resilience I urge all our congressmen and our legislators to pass that bill it's already a bill there's a bill on open data we need it for the pandemic we need it for all hazards in the country unless or in mitigate the harsh impacts of all hazards that are going to affect our country well, on behalf of the Filipino people I'd like to thank you guys academicians, mathematicians, scientists data scientists, disaster scientists for your contribution in this fight against pandemic I didn't realize it was so exciting to talk about statistician biologic mathematician and a disaster scientist geologist but I will just end this session by describing a quote from one of my professors in disaster medicine and he said today is the information age and we are swimming in oceans of data but the oceans of data need to flow to be organized and become rivers of information because if you organize data it becomes information you go to lakes of knowledge as you analyze the information you end up with knowledge but what is most important is said are the droplets of wisdom that you get from the lakes of knowledge with that, thank you very much for guesting my episode this evening and thank you everybody we do hope everyone has learned with our discussions on ncov.ph data management and science see you in the evening to all