How hacking can mitigate the impacts of natural hazards

Share on facebook
Share on twitter
Share on linkedin
Share on email

Niclas Rieger | Centre de Recerca Matemàtica

Nikos Mastrantonas | European Centre for Medium-Range Weather Forecasts

Recently, we had the opportunity to participate in the Copernicus Barcelona Hackathon, an intense 3-day hacking event (16-18 October 2020). We were fortunate to contribute in the development of ideas and prototypes that can support the mitigation of risks related to natural hazards. In this blogpost, we would like to present our personal experiences before and during the event, give some information about the projects we worked on, and briefly summarize why we really enjoyed our involvement in this hackathon.

But first, let’s give a short background about what a hackathon is and what it aims at…

A hackathon is an event that brings together participants of various fields that work together in interdisciplinary groups tackling a specific challenge within a short timeframe. The goal is to develop innovative ideas and translate them into prototypes with the delivery of a functioning software/hardware by the end of the event.

In this context, the Copernicus Hackathon Programme brings together developers, entrepreneurs and topic-specific experts to develop new applications based on Copernicus Earth observation data and services. Every year there are multiple hackathons across Europe organized by the programme, with the Copernicus Barcelona Hackathon being one of them. 

The specific challenges addressed by this event were related to natural hazards, a very relevant topic to the CAFE project and our professional interests. Also, due to the ongoing COVID pandemic the event was fully online, making it easier to participate, even from the comfort of our rooms 😊 Thus, we decided to attend the event and experience this format for the first time in our lives!

Niclas

Thursday, 14:00h (73h left to final pitch)

The team. More than 50 people unknown to me (except Nikos) cavort in the Slack Channel, which is supposed to be the organizational base of the Hackathon. It’s probably due to my (too often and too gladly taken) role as an observer that my team size so far has reliably not increased above 1. I aim somewhat uncertainly into the crowd and based on indicators that are important to me (my gut feeling), I write to a few participants directly. Lucky strike. Not least thanks to Elaine’s enthusiasm and positive energy, within a few hours I’m in Team TBD*, a colorful mixture of Business Development, GIS, Machine Learning and Climate Science.

 *TBD = to be decided; at this point I’d like to advise future Hackathon teams against this frivolous naming strategy, if you don’t want to find out during the final presentation that you never changed the name 😉

Friday, 19.00h (44h left to final pitch)

The idea. The Hackathon has now officially begun! Energy wave! Enthusiasm! Our team is ready to go, the discussion is in full swing! The numerous introductory presentations have shown: everything can work here. From relevant weather and climate data solutions to the combination of OpenStreetMapping and the integration of social media data. The range of possibilities is enormous … and overwhelming. Decide quickly now and then go!

Fig. 1 | The severity of Covid-19 cases is probably influenced, among other things, by the level of air pollution. Primarily an urban problem, air pollution is mainly driven by man-made emissions from traffic and industry. But also natural phenomena like wildfires or volcanoes can affect the air quality.

Saturday 10.00h (29h left to final pitch)

The (lack of)understanding. Shaking my head inwardly, but outwardly calm and composed, I declare for the umpteenth time: “The currently existing data only gives a certain resolution. Technically, it is not possible to go beyond”. On the other side sits Elaine who replies confidently: “I understand, but from the point of view of the business model, it doesn’t make sense. We will practically not be able to market this product. We need better resolution!” Inside, she’s probably shaking her head right now.

Sunday, 03.00 h (12h left to final pitch)

The pot of coffee. Empty. Again.

Sunday, 11:00 h (4h left to final pitch)

Yeah, sure. Only one hour left until the delivery of the prototype and the only thing we can get out of the Python script is a RuntimeError. A bit typical and definitely one of those classic moments where you curse Python, the computer and also yourself for doing this. After all, nobody forces us to do it, right? Well, focus! Time (as well as the computer) is against us, but to throw in on the last meters of the home stretch would be a bit silly. And not fair, after all, the rest of the team has the same situation.

Sunday, 17.00h (2h after the final pitch)

The home stretch. Somehow we have arrived. The analogy of the marathon is not so wrong. Despite the lack of sleep, the body mobilizes some energy reserves with a final shot of adrenaline (and some caffeine) in the last few meters. Relief as well as disbelief about what has just been achieved alternate with each other at an incredible pace. And while you somehow try to remember the individual stages of the last hours, you realize once again that together is easier than alone. 

In this sense, a big thank you to my team: Elaine, Mahabir and Yuan. You fought step by step until the end for the seemingly unreachable 1st place.

Fig. 2 | Our final product - 5-day risk assessment forecast of severe Covid-19 due to air pollution for France. We base our risk estimation on (i) the current number of Covid-19 infections (hazard), (ii) CAMS forecast of particulate matter <2.5 µm (PM2.5) (vulnerability) and (iii) the total population living in each French community (exposure).

Nikos

Joining a hackathon for the first time is already an interesting experience. Joining as an individual member without a preformed team and being unaware of the expertise, interests and the number of the other individual participants, has its additional challenges 😛 Nevertheless, uncertainties and the unknown are also part of life, and in fact a very exciting one. Thus, I decided to register for this event (thanks Florian for making me aware of it!).

The actual event started on Friday 16th October, but all participants were able to join a dedicated communication channel and start forming teams from Monday 12th October. And here starts the exciting journey of this hackathon and the roller-coaster of feelings and emotions.

To my surprise I noticed many familiar names among the participants; mainly friends and alumni from IHE Delft and the MSc I attended (Flood Risk Management), and co-volunteers at the Water Youth Network. A nice start to boost up the levels of confidence and enthusiasm. And now comes the step of joining a team: so many people with a lot of different experiences, expertise and backgrounds, certainly a nice way to form a complementary team, … or to get lost in the translation and the challenges of interdisciplinary collaborations. 

By Tuesday I had already joined a team aiming at providing solutions for users in Tanzania; an idea that I found very exciting and helpful for an audience with less direct access to information. Full of energy, together with the other teammates we started researching Copernicus products to be used and ways to convey the information to the end users. But smooth and straightforward pathways are not always feasible. Unfortunately, by Saturday noon two of the team members had to withdraw due to unexpected personal obligations, which made the completion of the project an impossible task. Therefore, we decided to work on this idea outside of the hackathon and abandon it for this event. Once again it all came back to the start, searching for a new team and project idea. But this time, already mid-way through the hackathon.

I was about to stop the quest, when two of my former IHE colleagues, Adele Young and Thaine H. Assumpção, informed me that I could join their team. And the excitement reaches the sky once more. The three of us, together with Gijs van den Dool, a Natural Catastrophe Modelling specialist, formed a team looking at ways to improve and update the provided information about hazard, vulnerability, exposure and risk. And the name of the team… Up2Date.

Fig. 3 | Up2Date team - from upper left clockwise: Gijs van den Dool, Adele Young, Thaine Assumpcao, Nikos Mastrantonas

Everyone started suggesting ways to make the idea more concrete and specific; advocating for particular products, tools, study areas, end-users, etc. By Saturday night, after hours of debating and productive discussions (in which at times temperaments flared, resulting in even better conclusions), we managed to finalize our idea. We focused on the agricultural sector and how we can use the latest available remote sensing information to improve the vulnerability assessment and increase the preparedness of the stakeholders. And so, here we are, tired and exhausted, working hard all Saturday night till Sunday morning, to finalize the required deliverables and preparing the slides for the final pitch. Lots of stress to finish on time, and excitement for pitching our idea to the evaluation committee and the rest of the participants. At this stage the support from the organizers and mentors was invaluable and special thanks should be given to Milan Kalas, Alexandre Sanchez and Calum Baugh.

Fig. 4 | Huge storm front approaches an agricultural field. Providing more accurate vulnerability information (e.g. crop phenology, real-time cultivated areas) can facilitate improved decision making.

Eventually hard work pays off, and we managed to get 2nd place at this event! 😊 Hackathon is already over, but our project idea and prototype are in their initial stages. Now we are searching for opportunities to further develop this work. Let’s see what the future will bring …

Concluding remarks

And what conclusions do we both draw at the end of the day (or better the weekend)? Perspectives, perspectives and perspectives.

(i) For us as young scientists, the continuous focus on the final product, to deliver quickly applicable results and all this under high time pressure was sometimes new – and therefore exciting! It may also have been a small advantage that we as CAFEtarians know how to deal with long working hours and few breaks (coffee!).

(ii) Complex problems can sometimes be solved in a certain way. But sometimes they can be better solved in a different way. Interdisciplinary groups allow us to tap the full potential of solutions and make possible what we as experts in our field can hardly do alone.

(iii) A more diverse group broadens the possible solution horizon, but also broadens the space for possible conflicts. And here too, perspective is important. Situations, discussions and problems can be perceived as tiring, frustrating and demotivating, but they can also be challenging, stimulating and guiding. In the end it is up to us to decide which perspective we want to take.

Therefore, we would like to encourage more people to attend such events, get out of their comfort zones, and live the experience of a hackathon; this is surely something we would attempt again in the future! Till that time, back to our daily life and the CAFE-related tasks! And in fact, there are a lot of interesting things happening in the CAFE world, so stay tuned for upcoming publications and outreach activities! 😊

—————————————————————————————————————————————

Niclas Rieger [ResearchGate profile] is an Early Career Researcher and PhD candidate at the Centre de Recerca Matemàtica (CRM). His research topic is the “Sub-seasonal forecasting of extreme precipitation events using sea surface salinity and other sea surface variables as predictors”.

Nikos Mastrantonas  [ResearchGate profile] is an Early Career Researcher at the European Centre for Medium-Range Weather Forecasts (ECMWF), and a PhD candidate at the Technical University Bergakademie Freiberg (TUBAF). His research topic is the “Predictability of large-scale atmospheric flow patterns over the Mediterranean connected to extreme weather”.

Share this post

Share on facebook
Share on twitter
Share on linkedin
Share on print
Share on email

More articles