Data Science Challenge
✳ Tue, 7th
☷ 13:30 - 15:00
☉ APB E007
Get on your bike and go!
A good bicycle infrastructure forms the basis for sustainable, environmentally friendly, and resource-saving mobility in urban areas. However, the traffic transformation is only progressing in very small steps. With our Data Science Challenge, we want to open up new perspectives and fields of action with the analysis of digital bicycle data.
Surely cyclists who are on the road every day know exactly the problem areas of cycling in their own city: How to avoid large, unsafe streets? Where does the bike lane suddenly end in a pothole? How to get from A to B with as much green view and as little bicycle and car traffic as possible? While this is the knowledge that each individual bicyclist carries within themselves, analyzing publicly available urban bicycle traffic data and other urban geographical data can help us connect individual perspectives to analyze the big picture of urban bicycle infrastructure.
Four teams have participated in the Data Science Challenge and will present their results that will be evaluated by a jury of experts from research, cities and industry.
DSC 1 Recommending Alternative Cycling Routes via Predicted Usage Patterns (Men, Dakai; Becktepe, Jannis; Esmailoghli, Mahdi; Bermbach, David; Abedjan, Ziawasch)
DSC 2 Analyzing Cargo Bike Usage in Leipzig for Improved Bike-Sharing System (Petersen, Hauke; Plank, Martin)
DSC 3 In-Database Machine Learning on Bicycle Data from Munich (Großmann, Christoph)
DSC 4 Predicting Bike Traffic Using Graph Neural Networks: Integrating Residential Density, Amenity Distribution, and Street Networks (Chou, Wen-Chuang)
Jury
Rico Herzog, City Science Lab der HafenCity Universität Hamburg
Konrad Krause, Geschäftsführer des ADFC Sachsen
Mirko Mühlpfort, Referat Digitale Stadt der Stadt Leipzig
Eric Peukert, Data Science Center Leipzig, ScaDS.AI
Award ceremony
The Award Session will be held during the Dresden City Hall Reception on Wednesday night. First to third place will be rewarded with prize money from a pool of 1000 euros.
Cooperation partners
The Data Science Challenge 2023 is carried out by ScaDS.AI Dresden/Leipzig in cooperation with BTW 2023 and the CUT (Connected Urban Twins) project of the City of Leipzig.
It is kindly supported by ↗ InfAI Leipzig.
The task for applicants was:
Choose a metropolitan area or a city with a sufficient density of sensors and other publicly available geodata for your analysis. In the linked document you can find ↗ exemplary data sources from Berlin, Munich, Hamburg, Dresden or Leipzig. Make sure that only trustworthy sensors are used, which do not provide false readings and have only a few dropouts in the recordings. If necessary, blend the data from different sources and clean them to ensure data quality.
Find interesting facts and patterns in the data sources and create an analysis that answers a question of social relevance. Possible initial approaches might include:
How has bicycle traffic changed over the last 5 years?
What factors influence measured bicycle traffic volumes? Can the volume be linked to specific events and thus predicted?
What factors and local conditions help predict bicycle traffic on specific street segments? Can the environment (e.g. the location of residential areas, shopping facilities, schools, train stations, bus stops, universities, and the structure of the road network, bicycle routes, district routes, and narrow sections at bridges) be used to infer where heavy bicycle traffic can be expected? Can a model be developed for this purpose that can be transferred to other cities?
Is there a correlation between population structure, election results, and other sociodemographic factors in city districts and the volume of bicycle traffic at specific measurement points?
Can correlations between measured bicycle traffic volumes and weather events, such as snow, rain, or heat, be demonstrated? Can recommendations for action be derived from this?
The result of the analysis can optionally be visualized or presented in purely textual form, for example as a recommendation for action. You decide which technologies, cloud services, and cloud technologies are used. The approach can integrate available services and tools or develop new ones.
Criteria for prototype/concept evaluation
Novelty and implementability of results: how significant is the potential for predicting bicycle traffic at previously unmeasured waypoints, transferability of predictive models, if applicable
Completeness/scope of results
Social relevance
Data visualization