Quarantining positively diagnosed individuals and tracing their recent contacts to isolate potentially infected has been the current strategy to fight the pandemic in the early stages. However, if number of infections escalates quickly and become too large to be widely tested or contact traced, nations are forced to go into lockdowns.

As we know, nation-wide lockdowns are not a sustainable solution. The economic impact due to this drastic measure is devastating.

We have identified the reasons for a pandemic to quickly escalate out of control  to be a lack of accurate insights on the risk of individual infections and lack of continuous and real-time information on virus hotspots. Only with accurate and real-time information, easily accessible for everyone, we  are empowered to act effectively and sustainably in the fight against COVID-19.

We were driven by the thought,

can we successfully tackle the pandemic without going into nation-wide lockdowns?


Oh yes, in this hackathon we found that we can.

Our solution empowers individuals with easily accessible, accurate and real-time information on  individual risk of infection and virus hotspots, by leveraging mobile technologies, health data and existing disease modelling research.

Based on our market research, we found the users would like to have a tool for individual risk prediction

Mobile App

We have developed a mobile app that utilises GPS and bluetooth. We have also created end points where we can fetch pre-existing medical conditions of the users from the public health data. The app also periodically sends survey questions to those users who have been flagged under high risk by the app based on the data collected and analysed. We only collect a minimum of data and have outlined the specifics in our detailed GDPR guidelines. Notifications and interactions will be context based and will consider the users surrounding and mood to display questions and updates.

We use AWS as a backend cloud service, because it enables quick prototyping and also large simple scaling and security.

Do you own an Android phone and want to get a feel for the app?Download it here

Don't forget to set permission for installing third party apps in your android settings

No Android device but still want to test?Play with the mock up here

Risk prediction

In order to compute the risk of an covid-19 infection we use state of the art mathematical modelling and machine learning techniques

  • Individual Medical Care
  • Population Medical Data
  • Infection Data
  • Mobility Information
  • Monte Carlo Simulation
  • Reinforcement Learning

We take public large scale data sources such as the development of the virus spread and general health information about the population provided by governments and institutions into account. Additionally we collect our own data via the surveys, the encounter and the movement analysis inside the app. With Monte Carlo Simulations we predict future virus spread and use this as in input for our Reinforcement Machine Learning Agent, who will be able to predict infection risks. This agent will be trained by validation data coming from test results and contact tracing solution efforts e.g. from Android & Apple or Pepp-PT.

We base this technology on existing research on Coronavirus spread models from earlier epidemics like MERS from leading researches in South Korea Kim et al. 2018. We adjust this model with novel ML techniques and more complex data, utilising the unique willingness of the population to fight this virus and their openness to help.

Updates about our implementation can be seen here

Business Model

We have a 2 sided solution with subscription model for users and with business and governments on the other side.

Our business research showed that the average user would be willing to pay 3.84€ per month for a solution like ours

So we decided to offer the service in the app store initially for about 4 euros. This is will allow us to grow our active user base. Once we have confidence in our model we will be able to sell our predictions to businesses and governments and this will allow us to break even and scale-up

Data Privacy Strategy

Data Privacy

We are aware that reservations against the extensive use of data in the fight against COVID-19 exist, especially in Europe. But, with the devastating effects of the shutdown growing every day, solutions are needed that use data and technology to overcome COVID-19 fast and efficiently. At the same time, we believe in data privacy as a cornerstone of any free and democratic society.

Our business model allows for Data Privacy Strategy based on full transparency and user control. We are not relying on the exploitation of personal data, so we do not need to hide what we do from our users. Adapting a forthcoming, highly transparent communication strategies, will allow us to build trust, stay compliant and benefit from strict data privacy rules. We will clearly explain what we are going to do with the users' personal data and obtain explicit permission from users for processing. Because health data is very sensitive for most users, we will give users maximum control over their data, beyond what is legally required.

Consent is the most transparent, user friendly and least risky justification for processing personal data, including health data. Consent allows us to process all the data we need, to provide the risk predictions. With regards to health data, which is specially protected under the GDPR, consent is also the most practical and fairest justification for processing.

Since giving consent to data processing is a standard procedure for most end user applications, asking our users for it, will not significantly lower conversation rates. We will use transparency and fairness as competitive advantages, by building data protection into the DNA of the product right from the start.

At every development step we consult with our team member and legal adviser Bork Morfaw, who is determined to to ensure GDPR compliance to the fullest extent for a functioning solution.

Further detailed information here in this document

Scale-up Plans

We have developed the solution foreseeing future pandemics with different behavior and  incubation period. We also plan to scale up the products to be a software solution for the public health domain across nations which can be used to easily reach out to its citizens for in-app surveys which is also valuable in containing seasonal epidemics. Many epidemiology models work similarly, that our app will be able to adjust to different outbreaks all over the world.

This can help preventing future pandemics and following global recessions.

Socio-economic Benefits

Our solution has various socio-economic benefits, especially when deployed widely. For example, we can help avoid shutdowns, advance European Sustainable Development Goals, support the fight against infection diseases, and support resource allocation by providing insights into the spread of the disease.

  • Schools and Universities could remain open
  • Remote work can be reduced as only high risk infectious people will stay home
  • Unemployment will be reduced due to higher mobility
  • Travelling will be possible, since our system can be used world wide
  • Resources can be allocated to upcoming hotspots of the epidemic
  • Faster response times and preparation leads to fewer deaths

Further detailed information here in this document

Progress made during the hackathon

During this hack we develop the mobile-app and the risk estimation model from the ground up. We came in as a team of 7 people from 4 countries, we met most of us at different hackathons competing against each other. We decided to join forces to inspire, innovate and engineer a solution that would save nations from going into lockdowns.

We started with digging deeper into the news, mentors and webinar to identify the most important problem or the pain point. Further, we spend ample time on brain storming ideas over discord channel which was active for past 60 hours without break. Once we had clear path to the solution we divided the tasks and started developing the solutions from scratch. We kept on updating the progress every 2 hours and made sure everyone is motivated and kept in the loop.

In short, we together developed a functional app from not knowing each other during the weekend.

Necessities in order to continue the project

PREDICT - 19 solution was developed as a stand alone fully functional service without any dependencies. Our app is able to make predictions and give real-time virus hotspots without dependencies. However, in order to get better risk predictions and more accurate virus hotspots, access to the pre-existing medical conditions of the users from the public health data is needed.  This is to be done with strong data protection agreements signed.

Feedbacks please

We would love your feedback to improve our hard work. Please download PREDICT-19 android app from link or from the Try it out section below. Looking forward to your valuable comments.  Remember to grant access to foreign apps while installing it.

Try It out



amazon-web-services, kotlin, node.js, python, swift, tableau, tensorflow

Devpost Software Identifier