2022 Cycling Hackathon: innovative mobility solutions for Belgium

On March 26 and 27, 2022, we hosted an online Hackathon with the European Cyclists’ Federation (ECF). With nearly 70 member organizations across more than 40 countries, ECF is the world’s largest and best-known cyclists’ advocacy organization. They aim to improve and increase cycling by influencing policy and harnessing the power of Europe’s cycling movements.

Throughout the weekend, six teams had the chance to learn and test new skills based on data analysis, visualization, and coding. It was an exciting, dynamic event with impressive solutions!

The challenge and data sets

We kicked off Saturday morning with an introduction to the challenge. Guillaume Carton, Data Scientist at Agilytic and the event’s organizer shared, “For this Hackathon, you should imagine yourselves in our Belgian Prime Minister’s shoes. The idea is that you want to improve bicycle use and infrastructures in Belgium. However, you have a limited budget. You want to focus first on the city/region where you’ll have the most significant impact.”

With data, the goal was to select which place is the best to invest in and promote bicycling. Then, explain why that is through visualization, data analysis, and business logic. The participant could take different angles to make their analysis. For example, to analyze the situation regarding the level of development (i.e., where infrastructure is lacking), the number of bicycle accidents, who uses current infrastructures and when they use them, the level of pollution in some areas, etc.

Each of the six teams received data sets related to cycling use and infrastructure to analyze and present in front of a jury at the end of the challenge. The data sets included open street map data, road accidents from 2019 to 2020, air pollution, Uber open data, and bike-sharing Villo stations. Teams were free to use other external data sources as well.

We encouraged teams to share their collaboration on LinkedIn, which you can see with the hashtag #ECFxAgilyticHackathon22.

Coaching sessions for Hackathon support

Policy Officer of Infrastructure Aleksander Buczyński supported the Hackathon as a coach and judge. He has a Master’s degree in Computer Science from the University of Warsaw. In the eight years before joining ECF, Aleksander worked at the General Directorate for National Roads and Motorways in Poland on cycling projects and road safety. “We like cycling, and it was nice to hear that the participants like it too. But we had to provide a critical look and ask – do the numbers really support the case?,” said Buczyński. Eleanor Denneman of ECF also joined as a judge and coach, she works as an Infrastructure Policy Intern at ECF and has a Master’s in Urban Studies from VUB.

Agilytic experts offered coaching sessions to help deepen participants’ understanding of data analysis. “As a coach, my role was mainly to listen to the participants’ solutions and help them clarify their ideas. When needed, I offered certain directions to solutions and encouraged teams to take the extra step that the jury would appreciate the most,” said Guillaume Lamine, Data Scientist at Agilytic.

The Hackathon’s final solutions

We asked each team to present their insights and methodology in 10 minutes maximum, in addition to 5 minutes of Q&A. It was a formidable competition with six teams of primarily master’s students and professionals. Here are the data challengers that stood out and took the top three places:

💡Solution 1 from Team 3 🥇

The 1st place team, with Roxani Gkavra and Pantelis Kaniouras, gave the best insights and presentation in terms of methodology, storytelling, visualizations, and delivery.

Julien Theys, Agilytic’s Managing Partner, congratulated the team on the award ceremony stage, “We saw really interesting, diverse, and creative approaches from all teams - whether looking at accidents, technologies, or locations. At the end of the day, Team 3’s project prevailed - showing a critical mindset, comprehensive index, and a very thorough and solid methodology.”

As the winners, Team 3 (pictured below, presenting their solution during the award ceremony) took home a cash prize of 1.000 €. Their presentation “A story of data and cycling potential in Belgium” identified the areas where infrastructure would be most beneficial by developing a cycling potential index.

Cycling potential index variables (full formula in the images below):

‍i : area of reference, corresponding to a zip code/Commune area in Belgium

‍C: Cycling Potential Index

‍PnA: Level of air pollution from non-automobile related pollutants

‍PAj: Level of air pollution from automobile-related pollutant j

‍A: Intensity of road accidents involving cyclists

‍a: Weight based on road accident type (adults vs. children and adolescents)

‍IF: Level of coverage by existing cycling infrastructure

‍PP: Population ratio

Team 3 presenting their solution during the award ceremony.

“It was a great experience, we were two but it worked out well. We took advantage of all the data sets that were provided to develop a cycling potential index. It’s quite simplified but takes into consideration the parameters that affect cycling. We created it from scratch with the data we had available, based on logic and literature to combine different data sources and create a single linear equation,” said Gkavra while presenting during the award ceremony.

“Also, we wanted to show that it’s not only about absolute numbers, we all had data about accidents but if there’s a policy that wants to appeal to more younger people to cycle, we have to give more attention to accidents involving children and adolescents. So we decided to give a higher weight to regions where there are more accidents involving this group,” remarked teammate Kaniouras.

Gkavra finished the short presentation by adding, “Our approach is very scalable and adaptive, and in the end, we gave an answer to the problem; not just a methodology and approach but we can make a decision to identify areas with the need for an intervention. We really looked at the bigger picture, especially when considering this is the Prime Minister’s decisions around cycling policy.”

Their approach resulted in a meaningful selection and comparison of areas (581 communes) on a national level based on available detail. Moreover, they processed Open Street Map (OSM) data to extract data cycling infrastructure shared space and cycleways. For the analysis of road accidents, they emphasized the impact of accidents on children and adolescent cyclists. Finally, this resulted in three areas with high potential for immediate action and cycling investment: Antwerp, Charleroi, and Liege.

Team 3’s coach Guillaume Lamine commented, “Team 3 had expertise in transportation planning, use of geospatial data, and data science in general. Moreover, they had a good level of professionalism. Next to these hard skills, you could also see that they were interested in the topic and well organized in their thinking. They proposed a unique cycling index and made real suggestions for new cycling roads.”

While the solution was strong, Team 3 also added limitations to their presentation, making their argument more well-rounded. These limitations included having air pollution and traffic accident data based on a single value and not on historical data, the actual cycling demand was not part of the index, the shared space infrastructure was not considered, and the index was limited to the zip code level.

“I think the solution is flexible enough to be refined so that policy-makers can use it at a larger scale in Belgium and Europe. They were able to combine the given data in a clever way. I also liked that they added recommendations and next steps to improve their index,” stated Alex Schouleur, judge and Senior Data Scientist at Agilytic.

💡Solution 2 from Team 1🥈

Team 1 - with Raphaël Widdershoven, Maarten Wens, Arnout Verboven, and Karl Polak - presented their “Cycling Investment Plan” on reshaping the future of cycling with the smarter use of data. They took the judging panel through the scope (e.g., target and context), pedelec network, regions to invest in, and solutions in their presentation. In today’s environment, they highlighted possible improvements such as infrastructure, technologies, policies, advertisement, and quality data.

Taking a distinctive turn with pedelecs analysis (i.e., using electric bikes with increased speed and distance) - they looked at how they could transform trips to-and-from cities and potential commuting to replace cars. From this, they focused on the city of Gent, comparing the high demand in traffic for cycling to a much higher demand for car driving. They suggested improving road infrastructure to adjust to more high-speed traffic, wider lanes, fewer intersections, better road quality, and different infrastructure to accommodate e-bikes.

Alex Schouleur, the team’s Hackathon coach, expressed, “I was pleasantly surprised by the inclusion of pedelecs (electric bikes) to increase the range of people to commute, it would be useful to integrate it in Belgium’s cycling infrastructure strategy. Additionally, they integrated external data from surveys to gather the cycling demand by area. They made very clear arguments with a beautiful presentation and strong delivery.”

💡Solution 3 from Team 6 🥉

Team 6 - with Elias Tesfaye Makonnen, Sihan Wang, Xinyuan Xing, and Kevin Albert -  highlighted the problems of emissions, sustainability, safety, limited resources, and political pressure. The objective of their presentation was to find what cities to invest in, including how to reduce accidents and emissions. Moreover, the team wanted to see how to improve the standard of living score in areas and decrease energy costs.

As for their method, they sought to recommend a list of cities to invest in based on accident analysis and pollution analysis. They looked at the highest and lowest-ranked cities for bicycle, car, and motorcycle use. Afterward, they investigated cities with elevated CO2, NO2, and PM2.5 pollution. Then, they used visualizations to make recommendations that encourage cycling in urban areas, reduce accidents, and reduce vehicle emissions.

In conclusion, they identified Antwerp and Gent as priority cities to invest in for enhanced safety measures. They identified Manhey and Gedinne as potential places to promote cycling to motorcyclists and car drivers. They flagged the municipalities of Hainaut, Antwerpen, and Tielt as needing to improve air quality efforts with improved cycling infrastructure and access.

Hackathon takeaways

Regarding his overall impressions, Aleksander from ECF said, “I was impressed by the variety of approaches presented by the different teams. While the winning team made the best and most systematic use of the datasets provided by Agilytic, other teams brought additional data to the table to make a convincing case, for example - investing in e-bikes.”

From our experience, hackathons continue to be a great way to quickly come up with solutions to present-day challenges while creating a supportive learning environment for participants to test ideas.

“A hackathon has the opportunity to offer to participants a real use case of the data science field. They compete with one another, which gives the best outcomes possible. All that while offering a fun time because it is still part of a game-like challenge - with no consequences if you fail,” said Guillaume Lamine.

“Participants have the opportunity to work as a team and learn from each other. A hackathon is also short and intense so participants can showcase what they are truly capable of under pressure,” added Alex Schouleur.

‍If you’re interested in joining our next event or hackathon, keep up with us on LinkedIn! And, if you’re looking to take the next step in your data career, check out our available job openings in data science and data engineering.