AEDES: A platform for dengue prediction using climate and health data for epidemic management

The UNICEF Venture Fund is featuring members of our A.I. & Data Science Cohort. In this interview, learn more about Cirrolytix, a Philippines-based startup developing a platform for dengue prediction using climate and health data for epidemic management. Their solution PROJECT AEDES has been recognized as a digital public good. With support from the Fund, the Cirrolytix team will test, train, and deploy models for dengue hotspot prediction. The team will also develop a risk framework to test and evaluate model performance. 

Tell us about your startup and what you are building. 

In 2017, CirroLytix was formed with a vision of democratizing access to data and AI solutions. In late 2018, while busy with commercial consulting and data science education, the team decided to pivot towards social good to develop use cases for the UN SDGs. It wouldn’t be until late 2019 when we finally got started with this, through the NASA Space Apps Challenge.  

One of the challenges in 2019 was “Smash Your SDGs” which called for solutions that combined Earth observations with other data to solve SDG-related issues such as health. At the time, the Philippines was in the middle of a 5-year epidemic of dengue. At that point we knew our path was clear: we were going to use data to fight dengue.  

We formed CirroLytix to create social impact through big data, which, especially in public health, can be essential for practitioners to deliver care and treatment more efficiently, more accurately, and more cost-effectively. Our mission is to help governments, non-profits, researchers, international agencies, and local organizations succeed at addressing collective social problems around the world by harnessing the full power of their data. Our ultimate contribution to society is in enabling social impact organizations to make a positive data-driven change to the lives they seek to uplift. 

Through Project AEDES, we aim to improve public health response against dengue in the Philippines by predicting dengue cases from climate and digital data and pinpointing possible hotspots from satellite data. 

How would you describe your solution to a non-technical person? 

We have built an automated information portal that correlates dengue cases and deaths with real-time data from climate, Google searches, and satellite maps, giving an advanced indicator of when dengue will emerge and potential dengue hotspot locations. This portal is accessible publicly, but is targeted toward public health and local government agencies to give them advanced notice of dengue outbreaks and help prioritize resources. 

1. First, precipitation and temperature climate create mosquito-breeding environments 

2. Mosquitoes spread and get infected by existing dengue cases, thereby spreading the disease 

3. New infections cause alarm which drives internet searches for dengue 

4. Dengue cases result in deaths 

5. Cases and deaths are reported to public health officials 

Therefore, by detecting A and C early, we hopefully disrupt the disease cycle before an epidemic spreads.  

What type of social issue/ impact area does your solution aim to solve? 

The main challenge we wanted to address was getting real-time dengue alerts since data took time for the Philippine health ministry to aggregate from the ground level. Our idea was, if there were a way to relate dengue to other data that was available on the fly, such as Google searches, weather, and satellite data, it would be possible to infer the ongoing dengue cases at any time. In addition, having access to NASA satellite data allowed us to pinpoint possible mosquito hotspots by detecting stagnant water from satellite readings. It was an end-to-end approach—we would predict dengue spikes and panics, and be able to tell health workers where to focus.   

What’s promising about the use of artificial intelligence and data science technology? 

Social impact spans many domains such as health, poverty, and human rights, but we believe that social problems today are inherently data problems. We see data as an enabler of insights, productivity, and automation, and artificial intelligence is the key to unlocking these benefits.  

What is unique about your solution and how is it different from what currently exists?  
Since it's predictive, getting up to the minute dengue nowcasting can help with the response time. Predicting dengue outbreaks can serve as a way to prioritize areas/interventions and manage resources, unlike today where we can't predict where dengue cases can occur. Since dengue data is delayed by 2-3 months with mostly reactive interventions, by the time these interventions happen, people have died. 

Why does being Open Source make your solution better? 

Project AEDES was the product of a global hackathon, and open-source is part of its inception. We view open-source as integral in empowering creative communities to help improve and share our solutions. We were bred by hackathons, and now we conduct our own hackathons to create more open-source solutions to social problems, and those solutions can be incubated to realize and scale social value. 

Tell us more about your team. What makes your team diverse? 

The team has a diverse set of skills but we are all data professionals working in the same organization. Our solution Project AEDES requires different specializations to handle its various use cases. The roles in our team include Data Scientist and Machine Learning Analyst, Data Engineer, Geospatial Analyst, Solution Architect, Community Manager, and external research collaborators. A diverse set of skills provides more avenues for ideas, innovation, and experimentation. 

While there’s a diversity of skills, the team has an equal mix of men and women working together to build a tech solution for dengue. We offer a flexible and fully virtual working environment that allows working mothers to participate. We have also worked with For The Women (FTW) Foundation to collaborate with women in tech to improve the prototype. 

Why is diversity important for your startup? How does it add value? 

Merit and evidence are important drivers of diversity when onboarding collaborators. The purpose is what defines our startup regardless of background, culture, and education, since data is problem-agnostic and digital skills can be acquired pretty much by anyone on the internet. We aim to reduce inequalities by starting it in our organization. Being a self-propelling consultant in our organization empowers individuals to drive their creativity to solve the problems in society. Flexible working hours, stackable mission-driven projects, and a fully virtual setup provide a conducive environment for consultants to create socially relevant solutions. “If you build a road, people start going onto it.” We’re helping others find the path to innovation. 

What do you plan on doing with UNICEF’s Venture Fund investment and how will you use that to leverage raising follow-on investment? 
The UNICEF Venture Fund investment gives us a massive boost in productionizing AEDES through the development of an interactive info site that provides dengue information. It also enables us to understand users better and consider their pain points, create reliable and tested models, and publish preprocessed insights to solve the lack of dengue data in the country. Once the product is up and running, we aim to target additional funding through UNICEF Growth Funding for us to: 1) apply the solution to other mosquito-borne diseases (such as Zika, Wolbachia, and West Nile virus), 2) scale it to other countries that are experiencing the same hardships, and 3) develop the community further in building this open-source solution that benefits people. 

What challenges are you currently facing in building your solution and/or startup? 
If we want to make it a global tool, we need to access more high-quality and granular global data. Another challenge is the on-the-ground validation to confirm water hotspots and outbreaks in the area which may require community drives and local engagement with stakeholders. Funding for innovation is also lacking in this space despite the global clamor for social impact and health-related initiatives as it remains very niche. Lastly, we need engagement with local champions to provide user stories about dengue interventions within their communities. 

How can others support you in working towards overcoming these challenges?  
AEDES was built with collaborators and researchers who are passionate about health, dengue, and data. Without them, we would not have succeeded in our ideas. Additionally, the UNICEF Venture Fund exposed us to open-source development and funding opportunities which solved the main concerns when we started. 

The solution can be improved if we could get access to high-quality, granular, and up-to-date weather or satellite data from global space agencies such as NASA, JAXA, or ESA. Since AEDES is an open-source solution, access to the network and opportunities for funding may require unique partners who believe in innovative solutions using novel technologies. Partnerships could be made with local city health champions to understand their pain points and how dengue has affected their citizens, discover how they integrate insights and interventions, and develop the tools with the use cases that they care about. 

Cirrolytix company profile here.