Hackathon for the COVID-19


This year bitsxlaMarató, a hackathon for  contributing to La Marató de TV3, a public Catalan charity effort, is organised by the Barcelona School of Informatics (FIB), Hackers@upc (organisers of HackUPC) and the Barcelona Supercomputing Center-Centro Nacional de Supercomputación (BSC), which come together to fight COVID-19. After a first edition focused on rare diseases, we are joining forces to stop the effects of COVID-19. Full of creativity, health and technology, where teachers, researchers and any professional from the fields of health and technology (but also from other areas!) will work as a team, for 3 days in a row, online. Together, they will develop and look for solutions to face all the challenges posed by COVID-19.

This Marathon concerns everyone!
Do not even think about missing it!

The Hackathon will include training talks that are a must to attend! Besides, we hope to have very interesting guided activities! During the last day of the Hackathon, the teams will make the presentations and demonstrations of the solutions as well as proposals for solutions obtained to the challenges set out the first day.

Now listen up! There will be (symbolic) prizes! In addition, a prize for the best project as well!, even though all the solutions and donations for the TV3 Marathon will be prizes by themselves for the millions of people affected! All of them will be grateful for it!


Yes indeed! We are also looking for a donation! Before, during and after the Hackathon our intention is to collaborate with the TV3 Marathon through the donation of participants, sponsors, and the entire UPC community and beyond!



Hey, challenges are ready! LISTEN UP! For the most computational ones you will be able to use Marenostrum from BSC and Power9 + GPU

In addition, if you feel like using a cluster of ARMs during bitsxlamarato, you can try to do this hands-on/hackathon sobre ARMs on 17 December. People in charge of this hands-on will leave accounts open for you to use during the bitsxlamarato!

Looking for similar patients: the AI Doctor House conquers severe COVID-19!

There is a pressing need by healthcare professionals to access information relevant to clinical practice in a more effective way. Over 80% of clinically relevant data is essentially unstructured, mainly images like MRI and clinical texts.

One of the challenges faced by doctors is finding patients and clinical cases that show particular similarities to a given case (similar symptoms, diagnosis, treatments, or other characteristics) amongst the rapidly growing amount of clinical records and medical publications and the complexity of the data. Detection of similarities among patients or groups of patients is key for evidence-based clinical practice, the selection of patients for clinical trials, prioritizing patients for vaccination and for understanding the variability in clinical outcomes.

From a COVID-19 point of view, AI tools should distinguish between patients with and with no risk of a severe outcome, so that clinicians could intervene promptly. Specifically, this task aims to promote the development of systems able to detect similarities among a collection of clinical case texts.

Proposing and collaborating: BSC.

the objective is to be able to compute and measure similarity between patients represented by their clinical case, that is, the text describing their medical condition, previous morbidities, medical tests and treatments performed, diagnosis or outcome. This very complex scenario can in principle be approached by a diversity of methodologies ranging from text similarity techniques used to detect plagiarism, clinical concept detection, or even more advanced semantic textual similarity strategies dealing with the meaning of natural language through AI.
Access to medically relevant information hidden in clinical texts is one of the principal challenges for healthcare professionals in the AI digital age. Questions such as which are the symptoms of patients with a worse outcome, given similar comorbidities, medications or procedures are very difficult to answer without systematically processing clinical texts. Even simpler, epidemiological questions like how many days have passed before COVID-19 symptoms started or if patients had travelled to certain geographical areas can only be answered efficiently by means of computational tools. Similarities between patients can aid prognosis, diagnosis and decision making, saving vital time to healthcare practitioners.
Logo BSC
Logo Plan TL

Covid-Tracking on Campus!

How could we know what close contacts are on a college campus once we are informed of a student with a positive PCR test/test who has been on the campus? How can we identify a student's position in the classroom to be able later to detect close contacts (less than X, usually 1.5 or 2 meters) in the classroom? Is it necessary to have identified the location in the classroom or could we carry out some different system run by proximity or geolocation within the classroom? Once we have been informed of a positive PCR test/test, the monitoring of how the person is feeling, etc., could be similar to that proposed at the challenge for primary care centres. In this particular case, scheduling a PCR test/test would not be required... Having a close contact detection system on campus could be a way to help CAPs and universities find close contacts on campus when the person with a positive PCR test/test is unaware of this contact (they do not interact with other students, teachers, and people around them). Another thing is how could we detect close contacts outside the classroom ... in the corridors, in the restaurant, outside the building on the same campus?
Therefore, the challenge is to find a system that can help position students in and out of the classroom, and to detect close contacts efficiently on a college campus.
Proposing and collaborating: FIB.
The Barcelona School of Informatics already has a positioning system for the classroom, where each student/teacher can register the row and column where a student is placed as well as save the information. However, positioning and inserting the information into the system does not always work well due to misinterpretations of positioning and errors on data entry, amongst others. Could any system be developed to prevent incorrect data entry? Could a system be created that would be able to provide the exact position within the classroom on a specific date? Could it be improved either with a kind of a local Covid-radar system or triangulation in the classroom? Alternatively, would it be better to have some other system? Could anything be done so the system becomes widespread in any campus/school?
Having a close contact monitoring system on campus could complement and assist in case tracking in the classroom as well as on campus in general. In short, to be able to deal better with the spread of the virus!
Logo FIB

Clinical-microbiological characterisation of SARS-CoV-2 infection in the paediatric age

Can we reduce the number of children (paediatric population) to whom the microbiological study, associated with additional health costs and usually annoying, should be performed? Defining a predictive value, grouped or individually, of the clinical signs and symptoms associated to the clinical suspicion of COVID-19 can help! Can you find the best mechanism (screening strategies and algorithms) to predict paediatric COVID-19? That is the challenge!

All children suspected to be infected by SARS-CoV-2 visited in any of the participating Primary Health-Care centres and Emergency Wards of the COPEDI-CAT project in Catalonia are included in the database. We collect demographic, epidemiological, clinical and microbiological diagnosis data from all these children.

The collection and analysis of this information are carried out from the already available COPEDI-CAT database on the REDCap © digital platform.

Proposing and collaborating: COPEDI-CAT Project, Vall d’Hebron, BIOCOM-SC

  • Use of AI, data science tools or create new algorithms to determine which symptoms or sets of symptoms are predictors of a positive case in children.
  • Application of AI techniques to detect COVID-19 symptom patterns in paediatric cases and, if possible, generate a score.
  • Exploration of ways of visualising the results that allow for a better understanding of the conclusions of the challenge.

  • The clinical signs and symptoms associated to respiratory virus infections are nonspecific in the paediatric age.

  • SARS-CoV-2 infection shows clinical signs and symptoms that are very similar to other seasonal respiratory viruses.

  • During the seasonal period (autumn-winter-spring) where the peak incidence of these respiratory viral infections is high, we need to determine a more accurate differential diagnosis taking into account the degree of transmissibility of SARS-CoV-2 infection and its impact on the household contacts (quarantine and confinement).

Logo Copedi-cat
Logo Vall d'Hebron
Logo biocomsc

Do you have COVID-19 cough? Help distinguishing which type of cough you have.

Cough is another feature in the symptoms of COVID-19. The sound of coughing can be unique when you want to distinguish it from another infectious or other process. At the UPC we want to help people find out what kind of cough they have and whether it is compatible with a SARS-cov-2 type infection. Do you accept the challenge?

Proposing and collaborating: CREB and CCD.

The ultimate goal is building a system that can tell whether the symptoms are SARS-cov-2 compatible, using information provided by the user that will include recorded cough audios. The proposed solution can include, but is not limited to, machine learning models or algorithms using signal processing techniques. Let your imagination fly free, build an app or any other idea that you envision. To work on this challenge you will have data, your imagination and GPUs from the BSC! 
Fast tests and pre-screening tools can be very helpful to contain COVID-19 outbreaks. In particular, a diagnostic tool for COVID-19 based on respiratory sounds and coughing test will be very desirable to test the population to identify possible outbreaks. Interacting with the cell phone and recording the audio anyone can be tested in a very flexible and cost- efficient way. Moreover, such test can be run in countries where PCR testing is too expensive, contributing to the decrease of the spread of the pandemic worldwide.
Logo B2SLab
Logo CCD


COVIDTracking-Primary Health Care Centres: Help the front line of medical care! Automatise the tracking and monitoring of contacts and asymptomatic cases!

The waves of COVID-19 are causing the overburdening of our Primary Health Care Centres (CAP). This overload is given not only by COVID-19 cases and the need to track and monitor their contacts, but also by the fact of having to keep taking proper care of "non-COVID" patients. Over the last few months, care for these patients, many of them chronic, has been reduced with the consequent worsening of their health.
In addition, there are cases of overlapping when monitoring the contacts of positive COVID-19 cases between the different actors (doctors, nurses, COVID managers of the CAP and trackers), which reduces efficiency and leads to confusion.
Could we create a system that helps better coordinate the tracking and monitoring of positive case contacts in order to free the CAPs from some of those tasks when the system becomes overburdened?
The system would have to be able to adapt to the number of cases so that it would either allow automatising some of the tasks in case the system became overloaded or it would enable it to continue with a more personalised monitoring if there was no such overload.
For instance:
  1. Could we automatise when contacting the close contacts of a COVID-19 case once the COVID manager enters the contact details of a Positive? In other words, could we somehow free our health system from informing close contacts that they are close contacts when necessary - due to overburdening triggered by a large amount of cases? Could we inform them that they have to be confined and that they have to be tested in a CAP as well as to let them know which CAP is the closest based on their address, and schedule a visit for them? Currently, a specific person makes the first contact, but it is not always possible if the system becomes overburdened.
  2. Could we automatise the monitoring of an asymptomatic case, after having a positive PCR test/test in order to know if it is still well? Can we automatise the fact that an asymptomatic case has already done the days of confinement required and so we can discharge them if they have not symptoms? Generally, we could coordinate all these efforts and free our CAPs so that they can focus mainly on COVID cases with symptoms or ... non-COVID patients! Only when there is a case with symptoms an alarm system could alert the doctors/nurses. Because all these things, the challenge is ... could we create a system that automatises the entire process or part of the process depending on the overburdening of the system, from the first contact entered by the COVID manager to the discharge of the asymptomatic cases? That could save energy for everyone!
Proposing and collaborating: Professionals from the Besòs EAP, FIB and HackersUPC.
The solution has to take into account communication systems that are pleasant and easy to use as well as suitable according to the situation and social context of the user. It should also need to be able to track for how long each COVID-19 patient has been non-asymptomatic. Also, pay attention! The system must have an alarm system in case there are symptoms and notify the most convenient CAP to manage it. Systems that help to find a previous appointment or even automatise it with the CAP that corresponds to you can help too! Moreover, the system must help the coordination of cases between the CAPs, since there are close contacts that are not allocated to the same CAP from which the person with a positive PCR test/test came from. Finally, it is important that the system allows the patient to be monitored at any time (that is, if the message has been sent requiring patient isolation, if they have been discharged, if someone has contacted them, etc.). Regarding communications, it will be extremely important to consider the fact that messages should be sent in different languages ​​(for example, we have a large Pakistani population, so it would be very useful to be able to send messages in Urdu).
This system must reduce the time devoted to asymptomatic cases, in addition to better coordinating all communications with patients and speeding up the communication process with close contacts, in case the number of cases overburdens the system. It may help that either the system has different levels that allow having steps automatised to some extent or that are carried out in a personalised way by the CAP. Even, if possible, the fact of arranging the day and time to do a PCR test/test in a CAP or suggested centre could save management time in CAPs due to COVID cases, and therefore it can improve the feeling of patients of being well cared for, since time will be devoted to those who need it most. Only in cases that display symptoms, a specific person will be in charge of monitoring the patient when the system becomes overburdened.
Logo CAP Besos
Logo FIB
Logo HackersUPC


Friday 18 December

05:00 p.m. - Hacker registration
06:00 p.m. - Opening ceremony
07:00 p.m. - Hacking starts
07:30 p.m. - Team building
08:00 p.m. - Talk on Twitch
09:00 p.m. - Talk on Twitch
10:00 p.m. - Talk on Twitch
11:00 p.m. - Let's play!
11:30 p.m. - See you tomorrow!

Saturday 19 December

09:00 a.m. - Good morning hackers!
10:00 a.m. - Talk on Twitch
11:00 a.m. - Special activity on discord
03:00 p.m. - Let's play!
11:00 p.m. - Let's play!
11:30 p.m. - See you tomorrow!

Sunday 20 December

09:00 a.m. - Good morning hackers!
12:00 p.m. - Hacking ends
12:30 p.m. - Demo time!
15:30 p.m. - Closing ceremony (*)
16:30 p.m. - See you next year! Thanks for everything!

(*) we will know the winners



Logo inLabFIB
Logo Everis
Logo Hipeac
Logo IthinkUPC


Logo Nostrum BioDiscovery
Logo LleidaHack
Logo FIBAlumni