Covid360

Algorithm conceived for early prevention and detection of users exposed to an infection risk. Based on data treatment carried out privately, a system that intends to get days ahead from any other.

Comment

The problem this project solves:

It breaks most of the chains of infection of COVID-19 using predictive intelligence with a learning curve through smartphones, interrelating applications from all over Europe, and fully respecting privacy by design

The solution we bring to the table:

Currently, there is no app or approach using predictive intelligence on users crossings. Some solutions proposed might seem similar, but their target are only patients that have tested positive, so they only reach 5-10 % of the infected patients and their chains of infection, since up to 90-95 % of the infected patients NEVER get tested. For instance, 55 % are straight asymptomatic. In these terms (without predictive intelligence) any application will be inefficient, and if the applications of different countries are not related to each other, they will not be effective either.

Privacy level:

It's not possible to have one app for the whole of Europe (for instance, some countries have different apps for their regions). In this way, all the clinical and private data must be stored on each regional or national health system databases related with their own app. I order to achieve an absolute respect for privacy by design, and to allow users crossings of different countries, crossings and viral loads should be stored in user's device and submit to a peer-to-peer data processing.

Our model fully respects the user privacy, since the data processing is peer-to-peer and it is capable of interrelating without storing them, and the crossing data between users don’t have to go through a server. Therefore, it will be able to interrelate also apps from all over the world, since each administration will be able to give instructions in their server regarding how an app from another system can interact with theirs.

The solution’s impact on the crisis:

Should this system prove its efficiency in a real environment, it will drastically reduce the number of infected people, thus reducing mortality due to the decrease in hospital overcrowding, potentially becoming an essential part of the overall solution. This can happen in Europe and be replicated all over the world, being especially useful in countries that have a network and smartphone implementation but lack of health services like those in Europe.

The necessities to continue the project:

Both political and health authorities must commit to this. To achieve that, we have already made contact and have entered on a final phase where is it being decided which line of action to follow (ours amongst the other options) with parties like the Secretary of State of Artificial Intelligence of the government of Spain, the Office of the Deputy Adviser on Digitalization of the Community of Madrid, and TicSalut, the entity responsible for coordinating the development of a new app for smartphones that will be deployed in Catalonia.

Once any of these administrations (or any other) decides to put it into operation, we only have to develop the algorithm and its features in the background of the "base app" that are already being used. We can do this straight away with our team or we can coordinate this with the development teams of the base apps that are currently available; for instance, in the case of Madrid and many other communities of Spain, apps that have been developed by a team of tech companies which we are already in contact with. Similarly, we are in contact with other teams. The moment the administration approves the deployment, it can be up and running in two weeks.

The value of this solution after the crisis:

It can assist in the prevention of any future outbreaks or waves due to the virus mutation.

The prototype:

It is an algorithmic upgrade for all the current apps so they can use predictive intelligence and interrelate between them —it cannot be visualized in a prototype— . To access the support system, which generates huge volumes of data (like the value of impregnation per second between users) based on sociology, demographic conditions, etc; you need a user login (www.covid360.info / www.covid360.es), which can be obtained by contacting us (we have attached some screenshots of the system, exportations, databases, etc.), and then you’ll find a system in which you can press “Play” (after uploading a database) and, based on several dynamic parameters, it exports metadata that tells us how the viral load is transferred in a “smart population” with its daily crossings (contacts). Based on this, we extract data regarding the evolution of the pandemic (exact to those existing right now) and how this influences the number of infected people, mortality, even immunity generation, % of infection, etc., and we extract baseline impregnation values like "user VS user" or "user VS geography", and the values when the user must start to impregnate other users, or the value when the user must be called to confinement/test.

Presentation video: https://www.youtube.com/watch?v=DfK1eL3H1hk&feature=youtu.be

Pitch video for #EUvsVirus: https://www.youtube.com/watch?v=tXO3X0bAYCs&feature=youtu.be

Special thanks to: Gastón Borysiuk, Santiago Martínez, Ana Mozo, Yolanda Mateos, Ramón Chorques, Carmen Juarez, Clementine Suraud, Salvador Bosch, Gabrielle De Val, Dani Díaz, Pedro Muñoz, Sergio Sena, Sebastian Mysakowski, Martin Chollat, Lina Nikolopoulou

Share this project:

Updates